Generating an output document based on a document template via a machine learned model

Abstract

A computing apparatus for generating an output document, comprising: one or more memories to store instructions; and one or more processors to execute the instructions to perform operations comprising: receiving a plurality of source documents, receiving an input associated with a request to generate the output document based on the plurality of source documents and a particular document template, and generating, via one or more machine-learned models, the output document having the particular document template.

Description

Technical Field

[0001] This disclosure generally relates to generating content based on source content selected (identified) by a user, via one or more machine learning models. For example, this disclosure relates to a method and computing apparatus for generating content (output document) by implementing a document template generated from source content by one or more machine learning models, thereby assisting users in managing, organizing, and creating content. Background Technology

[0002] Based on current computing systems, large language models (LLMs) are capable of interacting with text content. For example, a user can copy and paste content from a document into a chat box to query the LLM about that content. The LLM can then provide output about that content (e.g., a summary). Summary of the Invention

[0003] Various aspects and advantages of embodiments of this disclosure will be set forth in part in the description which follows, or may be learned from the description, or may be learned by practice of the exemplary embodiments.

[0004] In one or more example embodiments, a computing device is provided for generating, organizing, managing, and creating content. For example, the computing device includes: one or more memories configured to store instructions; and one or more processors configured to execute instructions to perform operations including: receiving a plurality of source documents; receiving input associated with a request to generate an output document based on the plurality of source documents and a specific document template; and generating the output document having the specific document template via one or more machine learning models.

[0005] In some implementations, these operations further include providing a graphical user interface for presentation on a display device, the graphical user interface being configured to receive selections of the plurality of source documents.

[0006] In some implementations, these operations further include: the one or more machine learning-based models determining the intent associated with the output document based on content associated with each of the plurality of source documents.

[0007] In some implementations, these operations further include: the one or more machine learning models determining the style associated with the output document based on the content of each of the plurality of source documents.

[0008] In some implementations, these operations further include: the one or more machine learning-based models determining the format associated with the output document based on the content of each of the plurality of source documents.

[0009] In some implementations, the process by which one or more machine learning models determine the format associated with the output document based on the content of each of the plurality of source documents includes identifying multiple headings for the corresponding sections of the output document.

[0010] In some implementations, these operations further include: providing a graphical user interface for rendering on a display device, the graphical user interface including a plurality of user interface elements, wherein each of the plurality of user interface elements corresponds to a different style that can be applied by the one or more machine learning models to generate the output document having the particular document template; and receiving a selection of one or more of the plurality of user interface elements, wherein generating the output document having the particular document template via the one or more machine learning models includes: applying one or more styles corresponding to the selection of the one or more user interface elements among the plurality of user interface elements.

[0011] In some implementations, these operations further include: providing a graphical user interface for presentation on a display device, the graphical user interface including a plurality of user interface elements, wherein each of the plurality of user interface elements corresponds to a different title that can be applied by the one or more machine learning models to generate portions of the output document having the particular document template; and receiving a selection of one or more of the plurality of user interface elements, wherein generating the output document having the particular document template via the one or more machine learning models includes: applying one or more titles corresponding to the selection of the one or more user interface elements to portions of the output document.

[0012] In some implementations, the computing device further includes one or more databases configured to store multiple generative machine learning models associated with multiple different document types, and these operations further include retrieving from the multiple generative machine learning models at least one generative machine learning model associated with a document type indicated by the common content of the multiple source documents.

[0013] In some implementations, these operations further include: receiving a plurality of training documents associated with a document type, wherein each of the plurality of training documents includes one or more common features associated with the particular document template; and training the one or more machine learning models to learn at least one of the intent, style, or format for the particular document template of the document type.

[0014] In one or more example embodiments, a computer-implemented method for organizing, managing, and creating content is provided. The computer-implemented method includes: receiving a plurality of source documents by a computing system comprising one or more processors and one or more machine learning-enabled models; receiving by the computing system input associated with a request to generate an output document based on the plurality of source documents and a specific document template; and generating the output document having the specific document template via the one or more machine learning-enabled models.

[0015] In some implementations, the method further includes: determining, based on the content of each of the plurality of source documents, at least one of style, intent, and format associated with the output document by the one or more machine learning-enabled models.

[0016] In some implementations, the method further includes: providing a graphical user interface for presentation on a display device of the computing system, the graphical user interface including a plurality of user interface elements, wherein each of the plurality of user interface elements corresponds to a style, intent, or format that can be applied by the one or more machine learning models to generate the output document having the particular document template; and receiving a selection of one or more of the plurality of user interface elements, wherein generating the output document having the particular document template via the one or more machine learning models includes: applying the style, intent, or format corresponding to the selection of the one or more user interface elements among the plurality of user interface elements.

[0017] In some implementations, the method further includes: storing in one or more databases multiple generative machine learning models associated with multiple different document types; and retrieving from the multiple generative machine learning models at least one generative machine learning model associated with a document type indicated by the common content of the multiple source documents.

[0018] In some implementations, the method further includes: receiving a plurality of training documents associated with a document type, wherein each of the plurality of training documents includes one or more common features associated with the particular document template; and training the one or more machine learning models to learn at least one of the intent, style, or format for the particular document template for the document type.

[0019] In one or more example embodiments, a computing device is provided for generating, organizing, managing, and creating content. For example, the computing device includes: one or more memories configured to store instructions; and one or more processors configured to execute instructions to perform operations including: receiving a selection of a plurality of sample documents; receiving input for generating a document template based on the selection of the plurality of sample documents; and, in response to receiving the input, implementing one or more machine learning models to generate the document template based on at least one of a style, format, or intent associated with the plurality of sample documents.

[0020] In some implementations, these operations further include: receiving multiple source documents, receiving further input associated with a request to generate an output document based on the multiple source documents and the document template, and generating the output document based on the document template via the one or more machine learning-based models.

[0021] In some implementations, these operations further include: the one or more machine learning models determining, based on the content and structure of each of the plurality of sample documents, at least one of the style, the format, and the intent associated with the plurality of sample documents.

[0022] In some implementations, these operations further include: providing a graphical user interface for rendering on a display device, the graphical user interface including a plurality of user interface elements, wherein at least one of the plurality of user interface elements is selectable to change at least one of the style, format, and intent associated with the plurality of sample documents as determined by the one or more machine learning models; and receiving a selection of the one or more user interface elements to change the at least one of the style, format, and intent associated with the plurality of sample documents as determined by the one or more machine learning models, wherein generating the output document according to the document template via the one or more machine learning models includes: applying the at least one of the style, format, and intent changed according to the selection of the one or more user interface elements.

[0023] In some implementations, the computing device further includes one or more databases configured to store multiple generative machine learning models associated with multiple different document types, and these operations further include retrieving from the multiple generative machine learning models at least one generative machine learning model associated with a document type indicated by the common content of the multiple sample documents.

[0024] In one or more example embodiments, a computer-implemented method for organizing, managing, and creating content is provided. The computer-implemented method includes: receiving, by a computing system including one or more processors, a selection of a plurality of sample documents; receiving, by the computing system, input for generating a document template based on the selection of the plurality of sample documents; and, in response to receiving the input, implementing one or more machine learning models of the computing system to generate the document template based on at least one of a style, format, or intent associated with the plurality of sample documents.

[0025] In one or more example embodiments, a computer-readable medium (e.g., a non-transitory computer-readable medium) is provided that stores instructions executable by one or more processors of a computing system. In some implementations, the computer-readable medium stores instructions that may include instructions for causing the one or more processors to perform one or more operations associated with any of the methods described herein (e.g., operation of the server computing system and / or operation of the computing device). As described herein, the computer-readable medium may store additional instructions to perform other aspects of the server computing system and computing device, and corresponding methods of operation.

[0026] These and other features, aspects, and advantages of the various embodiments of this disclosure will become better understood with reference to the following description, the accompanying drawings, and the appended claims. The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate exemplary embodiments of this disclosure and, together with the description, serve to explain the relevant principles. Attached Figure Description

[0027] Referring to the accompanying drawings, a detailed discussion of exemplary embodiments is set forth in this specification for those skilled in the art, in which:

[0028] Figures 1A to 1B An example system according to one or more example embodiments of the present disclosure is described;

[0029] Figure 2 A flowchart illustrating an exemplary, non-limiting computer-implemented method according to one or more example embodiments of the present disclosure is shown;

[0030] Figure 3 Example block diagrams depict a notebook application according to one or more example embodiments of the present disclosure;

[0031] Figures 4A to 4H An example user interface screen of a notebook application according to one or more example embodiments of the present disclosure is shown;

[0032] Figures 5A to 5BAn example user interface screen of a notebook application according to one or more example embodiments of the present disclosure is shown;

[0033] Figures 6A to 6B Further example user interface screens of a notebook application according to one or more example embodiments of the present disclosure are shown;

[0034] Figure 7 Example notebooks or projects that can be represented in a particular manner according to one or more example embodiments of this disclosure are shown;

[0035] Figures 8A to 8B A flowchart illustrating an exemplary, non-limiting computer-implemented method according to one or more example embodiments of the present disclosure is shown;

[0036] Figure 9 Example block diagrams depict document extractor applications according to one or more example embodiments of the present disclosure;

[0037] Figures 10A to 10E Example actions associated with a document extractor application according to one or more example embodiments of this disclosure are shown;

[0038] Figure 11A A block diagram of an example computing system according to one or more example embodiments of the present disclosure is depicted, the example computing system being used to organize, manage and create content by implementing one or more machine learning-based models for training content and / or source content selected by a user;

[0039] Figure 11B A block diagram of an example computing device according to one or more example embodiments of the present disclosure is depicted, the example computing device being used to organize, manage and create content by implementing one or more machine learning-based models for training content and / or source content selected by a user;

[0040] Figure 11C A block diagram of an example computing device according to one or more example embodiments of the present disclosure is depicted, the example computing device being used to organize, manage and create content by implementing one or more machine learning-based models for training content and / or source content selected by a user. Detailed Implementation

[0041] Reference will now be made to embodiments of the present disclosure, one or more examples of which are illustrated in the accompanying drawings, wherein similar reference numerals denote similar elements. Each example is provided by way of explanation of the present disclosure and is not intended to limit the present disclosure. In fact, it will be apparent to those skilled in the art that various modifications and alterations may be made to the present disclosure without departing from the scope or spirit thereof. For example, features shown or described as part of one embodiment may be used with another embodiment to produce yet another embodiment. Therefore, the present disclosure is intended to cover such modifications and alterations within the scope of the appended claims and their equivalents.

[0042] The terminology used herein is for describing exemplary embodiments and is not intended to limit and / or constrain this disclosure. Unless the context clearly indicates otherwise, the singular forms “a,” “an,” and “the” are intended to include the plural forms as well. In this disclosure, terms such as “comprising,” “having,” and “including” are used to specify features, numbers, steps, operations, elements, components, or combinations thereof, but do not preclude the presence or addition of one or more features, elements, steps, operations, components, or combinations thereof.

[0043] It will be understood that although the terms first, second, third, etc., may be used herein to describe various elements, these elements are not limited by these terms. Rather, these terms are used to distinguish one element from another. For example, without departing from the scope of this disclosure, a first element may be referred to as a second element, and a second element may be referred to as a first element.

[0044] The term “and / or” includes a combination of multiple related listed items or any one of multiple related listed items. For example, the scope of the expression or phrase “A and / or B” includes item “A”, item “B”, and a combination of items “A and B”.

[0045] Furthermore, the scope of the expression or phrase “at least one of A or B” is intended to include all of the following: (1) at least one A, (2) at least one B, and (3) at least one A and at least one B. Similarly, the scope of the expression or phrase “at least one of A, B or C” is intended to include all of the following: (1) at least one A, (2) at least one B, (3) at least one C, (4) at least one A and at least one B, (5) at least one A and at least one C, (6) at least one B and at least one C, and (7) at least one A, at least one B and at least one C.

[0046] Large Language Models (LLMs) can generate general text content. However, if a user wants to generate content of a specific type or style using an LLM, the user either needs to experiment with multiple different prompting strategies or retrain (fine-tune) the LLM on multiple training samples that demonstrate that specific type or style of content. Experimenting with different prompting strategies can lead to unnecessary and redundant processing / content generation, as the user will try multiple times to create the desired content until the appropriate style is achieved. Retraining the model is computationally very expensive. Therefore, both approaches result in a waste of computational resources.

[0047] According to examples in this disclosure, a computational system is configured to automatically extract types or styles from one or more source documents (e.g., notes), and then apply the extracted types or styles to generate a template for creating a new (output) document. For example, the computational system can be trained to learn a specific template based on multiple documents of a specific type. As an example, a user can upload multiple Product Requirements Documents (PRDs), and the computational system can be configured to learn document structure and style. Subsequently, the user (or another user) can upload multiple source documents (e.g., multiple User Experience Research (UXR) documents) and submit a request to generate a document with a specific format (e.g., PRD format). The computational system can be configured to generate the PRD based on the multiple source documents and the learned document structure and style.

[0048] In some implementations, a graphical user interface (GUI) can be provided to the user to change or modify the style, format, and / or intent of a document template. For example, if the document template has a format including a title section, a description section, a background section, and a conclusion section, provided in a specific order, a GUI can be provided to the user, including multiple user interface elements that can be selected to modify the order or arrangement of these sections. Similarly, if the document template has a specific style (e.g., a clear viewpoint), a GUI can be provided to the user, including multiple user interface elements that can be selected to change that style from the document template to another style (e.g., a casual language).

[0049] In some implementations, the computational system can be configured to train one or more machine learning models to learn attributes of multiple training documents, such as document type and / or the intent, style, and format of the multiple training documents. The multiple training documents may share one or more common features (e.g., the same or similar document type, the same or similar document structure, the same or similar style, etc.).

[0050] In some implementations, the computational system includes one or more databases configured to store multiple generative machine learning models, each associated with a plurality of different document types. The computational system can be configured to retrieve from these multiple generative machine learning models at least one model associated with a document type indicated by the common content of the multiple source documents. For example, the multiple source documents may each contain content associated with a specific type of document (e.g., resume, PRD, research paper, etc.).

[0051] Under current computing systems, large language models (LLMs) are capable of interacting with text content. However, current computing systems require significant effort to create specific prompts for the LLM to process. For example, a user might need to copy and paste content from a document into a chat box to query the LLM about that content. This switching between multiple windows or applications results in a significant waste of computation time and resources (e.g., processor cycles).

[0052] According to examples of this disclosure, a computing system (computing platform, computing device) is configured to create new types of output (e.g., outlines, reports, summaries, etc.) based on source content provided to the computing system (e.g., by a user) via one or more machine learning models. For example, the computing system may be configured to receive source content selected by a user and generate a summary of the source content via one or more machine learning models, including identifying one or more topics related to the source content.

[0053] As an example, a user can identify and select a subset of documents (e.g., four documents) from multiple documents (a large document corpus) related to a specific topic (e.g., modern American history in the 1990s), which are then fed to the computational system. The computational system may include one or more machine learning-enabled models configured to receive the selected documents as input and provide summaries (or reports, papers, outlines, etc.) related to the selected documents, as well as identification of key topics (e.g., via document guides).

[0054] In some implementations, the computing system is configured to implement semantic retrieval methods (e.g., clustering) and one or more machine learning models (e.g., one or more LLMs) to generate summaries, key topics, and suggested queries (e.g., questions), thereby generating document guides for content identified or indicated by the user (e.g., based on the text body found in the content).

[0055] For example, in some implementations, the computing system can be configured to receive source content from a user. For instance, a user can upload source content (e.g., documents, image content, audio files, websites, videos, presentations, PDFs, etc.). In some implementations, the computing system can be configured to, in response to a user uploading source content, automatically generate information including a summary of the source content, popular topics found within the source content, suggested topics and questions to help the user further explore the source content, and so on. This information can be presented via a user interface. The user interface can be configured to receive input from the user (e.g., via touch input, mouse clicks, etc.) on user interface elements corresponding to topics, questions, etc. In response to receiving input from the user, the computing system can be configured to respond to that input, for example, by providing an answer to a query for a question or topic based on the source content via one or more machine learning models.

[0056] In some implementations, the computing system can be configured to automatically generate information, including reports, outlines, or rewrites of the original content, in response to user-uploaded source content, to generate new content based on the source content identified (selected) by the user. For example, a user could request the computing system to identify a specified number of topics from one or more documents, summarize customer interactions that occurred over a specified period (e.g., the last two weeks), generate a specified number of ideas based on the source documents, etc.

[0057] In some implementations, the source content that the LLM depends on or references can be selected by the user (e.g., curated). For example, the user can consider or indicate that the selected source content is trustworthy (e.g., trusted source content, authoritative source content, etc.) or has a higher priority compared to other content without such designation. Therefore, the one or more machine learning-enabled models are configured to generate content summaries or new content based on trusted source content, improving the accuracy and reliability of the information and data provided to the user. Furthermore, the one or more machine learning-enabled models are configured to answer questions about the trusted source content based on that content, improving the accuracy and reliability of the information and data provided as answers to user-posed questions.

[0058] In some implementations, the computing system can be configured to discover, add, or remove source content. For example, a user can add or remove source content. For instance, a user can provide input requesting the computing system to discover source content (e.g., by searching for academic articles on a specific topic), and the user can add the discovered source content as part of the selected source content that the user (and / or the computing system) deems trustworthy.

[0059] In some implementations, the computing system can be configured to receive additional source content by means of users creating new notes, users uploading the source content to the computing system, or adding the source content via a website. The computing system can be configured to generate or receive metadata about the added source content. For example, this metadata may include one or more of the following: title, author, upload date, date associated with the creation of the source content, and a Uniform Resource Locator (URL) associated with the source content.

[0060] In some implementations, the computing system can be configured to delete or remove source content by having the user select the source content and provide input requesting its deletion (e.g., from a notepad application). In some implementations, the source content may be deleted as a source relied upon when generating summaries, key topics, etc., in the notepad application, but the original copy of the source content may be maintained elsewhere.

[0061] In some implementations, the computing system can be configured to receive user input via text input boxes (e.g., open text input boxes). For example, the user input could be in the form of a question (e.g., "What did Nixon say in his speech about the use of automobiles?"). Alternatively, the user input could be in the form of a topic or idea (e.g., "Nixon's automobile crisis" or "What is this document about?").

[0062] The computational system can be configured to provide a response to user input based on selected source content via one or more machine learning models. In some implementations, the computational system is configured to indicate the number of sources (references) relied upon when providing the response. In some implementations, the computational system is configured to provide sources (references) relied upon by a specific passage in the response for presentation. In some implementations, the computational system is configured to provide additional context about the sources (references) relied upon by a specific passage in the response. For example, the computational system can indicate from the source the passage (e.g., a sentence or paragraph) on which a part of the response is based, and can further indicate from the source a preceding and / or following passage to provide further context about that specific passage.

[0063] In some implementations, the computing system can be configured to store one or more text segments (e.g., fragments) generated from responses (answers) to user-input queries (questions). For example, these segments can be stored in a designated area of ​​a notepad application. This designated area can be called a scratchpad, and each piece of information stored in the scratchpad can be called a note. The user can select one or more segments to store as a first note in the scratchpad. In some implementations, quotations can be stored as second notes in the scratchpad. In some implementations, the user can select (e.g., highlight) specific segments from quotations (source content) to store as third notes in the scratchpad. In some implementations, the user can store their own segments or comments as written notes (fourth notes).

[0064] According to examples of this disclosure, the computing system can be configured to provide a notepad application configured to generate output (e.g., outlines, reports, summaries, etc.) based on source content provided to the notepad application (e.g., by a user) via one or more machine learning models. The notepad application can be configured to allow users to create various projects to complete various tasks. Each project can be configured to function like a folder, allowing users to store various information into each project. In some implementations, a separate draft area may correspond to or be dedicated to a specific project. In some implementations, the notepad application can be configured to receive the source content specified by the user. The notepad application can be configured to add, delete, or modify projects based on input received from the user. Each project can be provided with a default name, a name provided by the user, or a name generated by the notepad application (e.g., via one or more machine learning models) based on information stored in that project (e.g., based on the source content).

[0065] In some implementations, in response to source content being provided to the notebook application, the notebook application can be configured to automatically generate (e.g., using one or more machine learning models, one or more generative machine learning models, etc.) a graphical image (e.g., an emoji, icon, etc.) or graphical animation corresponding to or representing the source content. In some implementations, the graphical image or graphical animation can be overlaid on a folder provided as a user interface element, which, when selected, causes the folder to open and display its contents to the user. Alternatively, in some implementations, in response to the source content being provided to the notebook application, the notebook application can be configured to automatically generate (e.g., using one or more machine learning models, one or more generative machine learning models, etc.) a text description (name) corresponding to or representing the source content. This text description can be overlaid on a folder provided as a user interface element, which, when selected, causes the folder to open and display its contents to the user.

[0066] One or more technical benefits of this disclosure include generating content based on specific content items selected by the user, via one or more machine learning models. Current methods for generating output for Large Language Models (LLMs) require the user to copy and paste content from a document into a chat box to query the LLM for that content. Switching between multiple windows or applications results in a significant waste of computation time and resources (e.g., processor cycles). In contrast to current methods, in response to a user-uploaded content item, a summary of the user-selected content item (e.g., the source content) can be automatically generated via a notebook application and one or more machine learning models. Therefore, the user does not need to switch between applications or windows, nor are prompts required.

[0067] Another technical benefit of this disclosure includes one or more machine learning models providing suggested queries based on content items selected by the user, suggested key topics based on content items selected by the user, and selectable tags based on the response output. The user can select a suggested query, and the one or more machine learning models can be configured to provide a response to that query based on the content item selected by the user. Automating the provision of the suggested query saves computational resources (e.g., network resources including bandwidth, processor cycles, etc.) by not requiring the user to input the suggested query.

[0068] Another technical benefit of this disclosure includes one or more machine learning models generating content based on content items selected by the user and / or notes selected by the user. Content generation can save time and computational resources because it does not require the user to cut and paste content from multiple sources to generate new content based on multiple content items (e.g., outlines, articles, reports, etc.).

[0069] Another technical benefit of this disclosure includes one or more machine learning models generating graphical images or animations to be displayed overlays on folders, thereby indicating the contents stored in those folders. These graphical images or animations can improve search capabilities and save computational resources that might otherwise be consumed by users opening and closing folders that do not contain the content they are actually looking for.

[0070] Another technical benefit of this disclosure includes one or more machine learning models providing suggested queries based on content items selected by the user, suggested key topics based on content items selected by the user, and selectable tags based on the response output. The user can select a suggested query, and the one or more machine learning models can be configured to provide a response to that query based on the content item selected by the user. Automating the provision of the suggested query saves computational resources (e.g., network resources including bandwidth, processor cycles, etc.) by not requiring the user to input the suggested query.

[0071] Another technical benefit of this disclosure includes one or more machine learning models generating document templates based on training documents that can be provided by a user. For example, a user may be unfamiliar with the typical (accepted) or standard format of certain document types (e.g., resumes, PRDs, legal opinions, etc.) and may lack the knowledge or expertise to provide appropriate prompts to generate a document template with the correct format for that specific document type. Therefore, the document extractor application described herein can improve the efficiency of the one or more machine learning models by generating document templates with appropriate structure (e.g., style, intent, format, etc.) that can be used to generate output documents based on the correct templates. Thus, reducing interaction with the user to generate the document template can save or conserve bandwidth, network resources, computing power, etc. Furthermore, reducing the inference performed by the one or more machine learning models to generate the document template can also save or conserve bandwidth, network resources, computing power, etc.

[0072] Another technical benefit of this disclosure includes one or more machine learning models generating output documents based on source documents that can be provided by a user and document templates that can be selected by the user or the computing device. For example, a user may be unfamiliar with the typical (accepted) or standard format of certain document types (e.g., resumes, PRDs, legal opinions, etc.) and may lack the knowledge or expertise to provide appropriate prompts to generate an output document with the correct format for a particular document type. Therefore, the document extractor application described herein can improve the efficiency of the one or more machine learning models by generating output documents with appropriate structure (e.g., style, intent, format, etc.) based on document templates suitable for that document type. Thus, reducing the interaction with the user to generate output documents via one or more machine learning models based on the generated document templates can save or conserve bandwidth, network resources, computing power, etc. Furthermore, reducing the inference performed to generate the output document by the one or more machine learning models can also save or conserve bandwidth, network resources, computing power, etc.

[0073] Now refer to the attached diagram, Figure 1A This is an example system based on one or more example embodiments of this disclosure. Figure 1A An example of system 1000 is shown, which includes computing device 100, external computing device 200, server computing system 300, and external content 500, which can communicate with each other via network 400. For example, computing device 100 and external computing device 200 may include any of a personal computer, smartphone, tablet computer, laptop computer, GPS device, smartwatch, etc. Network 400 may include any type of communication network, including wired or wireless networks or combinations thereof. Network 400 may include a local area network (LAN), wireless local area network (WLAN), wide area network (WAN), personal area network (PAN), virtual private network (VPN), etc. For example, wireless communication between the elements of the example embodiment may be performed via wireless LAN, Wi-Fi, Bluetooth, ZigBee, Wi-Fi Direct (WFD), ultra-wideband (UWB), Infrared Data Association (IrDA), Bluetooth Low Energy (BLE), Near Field Communication (NFC), radio frequency (RF) signals, etc. For example, wired communication between the elements of the example embodiment may be performed via twisted pair, coaxial cable, fiber optic cable, Ethernet cable, etc. Communication over the network 400 can use a variety of communication protocols (e.g., TCP / IP, HTTP, SMTP, FTP), encoding or format (e.g., HTML, XML), and / or protection schemes (e.g., VPN, Secure HTTP, SSL).

[0074] As will be explained in more detail below, in some implementations, computing device 100 and / or server computing system 300 may form part of an application system that can provide users with tools for creating, managing or organizing information (e.g., documents, image content, etc.) via one or more machine learning-based models.

[0075] In some example embodiments, server computing system 300 may obtain data from one or more of content data repository 350, user data repository 360, and machine learning-enabled model data repository 370 to implement various operations and aspects of the application system disclosed herein. Content data repository 350, user data repository 360, and machine learning-enabled model data repository 370 may be provided integrally with server computing system 300 (e.g., as part of one or more memory devices 320 of server computing system 300) or may be set up separately (e.g., remotely). Furthermore, content data repository 350, user data repository 360, and machine learning-enabled model data repository 370 may be combined into a single data repository (database), or may include multiple corresponding data repositories. Data stored in one data repository (e.g., content data repository 350) may overlap with some data stored in another data repository (e.g., user data repository 360). In some implementations, one data repository (e.g., machine learning-enabled model data repository 370) may reference data stored in another data repository (e.g., user data repository 360).

[0076] In some examples, the content data repository 350 can store any type of information or content. For instance, the content data repository 350 may include books, product manuals, resumes, legal opinions, academic papers, proprietary data files, patent documents, web pages, emails, forum posts, social media posts, videos, images, geographic information, or any other type or manner of content that can be stored or accessed digitally (e.g., in a database, storage device, etc.). In some implementations, information can be stored in the content data repository 350 by the user selecting certain documents, images, or other content to be stored there.

[0077] In some examples, the user data repository 360 may include information about one or more user profiles, including various types of user data such as user preference data, user demographic data, user calendar data, user social network data, user historical travel data, etc. For example, the user data repository 360 may include, but is not limited to: email data including text content, images, email-linked calendar information, or contact information; social media data including comments, ratings, check-ins, likes, invitations, contacts, or bookings; calendar application data including dates, times, events, descriptions, or other content; virtual wallet data including purchases, e-tickets, coupons, or promotional offers; scheduling data; location data; SMS data; or other suitable data associated with the user's account. According to one or more examples of this disclosure, the data can be analyzed to determine user preferences for generating, managing, and / or organizing content, such as automatically generating document summaries in a specific manner or style, automatically providing customized features for content, providing suggestions, recommendations, and / or questions related to certain content identified by the user as the source content, automatically generating document templates, and automatically generating documents based on those templates, etc.

[0078] The user data repository 360 is provided to illustrate potential data that, in some embodiments, may be analyzed by computing device 100 and / or server computing system 300 to identify user preferences, make recommendations, generate, manage, and / or organize content. However, such user data may not be collected, used, or analyzed unless the user has given consent after being informed of what data is being collected and how it will be used. Furthermore, in some embodiments, tools may be provided to the user (e.g., in a notebook application or via a user account) to revoke or modify the scope of permissions. Additionally, certain information or data may be processed in one or more ways before being stored or used, such that personally identifiable information is removed or stored in encrypted form. Therefore, based on the permissions granted by the user, computing device 100 and / or server computing system 300 may or may not be able to access specific user information stored in the user data repository 360, or such data may not be stored in the user data repository 360 at all.

[0079] The machine learning model data repository 370 can store machine learning models that can be retrieved and implemented by the server computing system 300 to generate distilled or fine-tuned machine learning models (e.g., distilled or fine-tuned generative machine learning models). In some implementations, these models can also be provided to the computing device 100. The machine learning model data repository 370 can also store distilled or fine-tuned machine learning models (e.g., distilled or fine-tuned generative machine learning models) that can be retrieved and implemented by the computing device 100. In some implementations, the computing device 100 can retrieve and implement machine learning models as large-parameter models that have not yet been fine-tuned or distilled. The machine learning models stored in the machine learning model data repository 370 (including large parameter models and distilled or fine-tuned models) can include generative machine learning models associated with different types of content (e.g., different genres or topics, different kinds of content (including image content, video, and text), different types of content or documents (e.g., outlines, reports, spreadsheets, etc.), documents with different styles (e.g., casual, opinionated, expert, etc.), documents with different formats (e.g., outline format, PRD format, resume format, etc.), and documents with different intentions (e.g., informing, persuading, comparing, contrasting, etc.). Machine learning models can include large language models (e.g., large language models with bidirectional encoder representations (BERT) from Transformers) and general multimodal models (e.g., Gemini). Machine learning models can include generative artificial intelligence (AI) models (e.g., Bard) that can implement generative adversarial networks (GANs), Transformers, variational autoencoders (VAEs), neural radiation fields (NeRFs), etc.

[0080] External content 500 can be any form of external content, including news articles, web pages, video files, audio files, written descriptions, ratings, game content, social media content, photos, business offers, modes of transportation, weather conditions, sensor data obtained from various sensors, or other suitable external content. Computing device 100, external computing device 200, and server computing system 300 can access external content 500 via network 400. External content 500 can be searched by computing device 100, external computing device 200, and server computing system 300 using known search methods, and search results can be ranked based on relevance, popularity, or other suitable attributes (including location-specific filtering or promotion).

[0081] Now for reference Figure 1BThis will describe example block diagrams of a system 1000' including a computing device 100 and a server computing system 300 according to one or more example embodiments of the present disclosure. Although the computing device 100 is... Figure 1B The features of the computing device 100 described herein are also applicable to the external computing device 200.

[0082] The computing device 100 may include one or more processors 110, one or more memory devices 120, an application system 130, a location determination device 140, an input device 150, a display device 160, an output device 170, and a capture device 180. The server computing system 300 may include one or more processors 310, one or more memory devices 320, and an application system 330.

[0083] For example, the one or more processors 110, 310 can be any suitable processing device that can be included in computing device 100 or server computing system 300. For example, the one or more processors 110, 310 can include one or more of the following: processor, processor core, controller and arithmetic logic unit, central processing unit (CPU), graphics processing unit (GPU), digital signal processor (DSP), image processor, microcomputer, field-programmable array, programmable logic unit, application-specific integrated circuit (ASIC), microprocessor, microcontroller, etc., and combinations thereof, including any other means capable of responding to and executing instructions in a defined manner. The one or more processors 110, 310 can be a single processor or multiple processors operatively connected, such as in parallel.

[0084] The one or more memory devices 120, 320 may include one or more non-transitory computer-readable storage media, including read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), and flash memory, USB drives, volatile memory devices including random access memory (RAM), hard disks, floppy disks, Blu-ray discs, or optical media such as CD-ROMs and DVDs, and combinations thereof. However, examples of the one or more memory devices 120, 320 are not limited to those described above, and the one or more memory devices 120, 320 may be implemented by various other means and structures that will be understood by those skilled in the art.

[0085] For example, the one or more memory devices 120 may also include data 122 and instructions 124, which may be retrieved, manipulated, created, or stored by the one or more processors 110. In some example embodiments, such data may be accessed and used as input to implement a notebook application 132 and execute instructions to perform operations, including providing a user interface comprising a first portion and a second portion, wherein the first portion includes a text summary generated by one or more machine learning models based on a plurality of documents selected by a user, and the second portion includes a plurality of user interface elements for performing operations on the text summary, as described in the examples of this disclosure. In some example embodiments, such data may be accessed and used as input to implement a document extractor application 136 and execute instructions to perform operations, including receiving a plurality of source documents, receiving input associated with a request to generate an output document based on the plurality of source documents and a specific document template, and generating the output document having the specific document template via one or more machine learning models, as described in the examples of this disclosure.

[0086] For example, the one or more memory devices 320 may also include data 322 and instructions 324, which may be retrieved, manipulated, created, or stored by the one or more processors 310. In some example embodiments, such data may be accessed and used as input to implement a notebook application 332 and execute instructions to perform operations including providing a user interface comprising a first portion and a second portion, wherein the first portion includes a text summary generated by one or more machine learning models based on a plurality of documents selected by a user, and the second portion includes a plurality of user interface elements for performing operations on the text summary, as described in the examples of this disclosure. In some example embodiments, such data may be accessed and used as input to implement a document extractor application 336 and execute instructions to perform operations including receiving a plurality of source documents, receiving input associated with a request to generate an output document based on the plurality of source documents and a specific document template, and generating the output document having the specific document template via one or more machine learning models, as described in the examples of this disclosure.

[0087] In some example embodiments, computing device 100 includes application system 130. For example, application system 130 may include notebook application 132, document application 134 (e.g., word processing application, spreadsheet application, presentation application, image content application, etc.), and document extractor application 136. Application system 130 may include various other applications, including text messaging application, email application, dictation application, virtual keyboard application, browser application, map application, social media application, navigation application, etc.

[0088] According to examples of this disclosure, notebook application 132 can be executed by computing device 100 to provide a user of computing device 100 with a way to organize, manage, create, and interact with content, particularly content selected or chosen by the user. In some implementations, notebook application 132 may be part of document application 134 or may be a standalone application. Notebook application 132 can be configured to dynamically interact based on various user inputs. Example implementations of notebook application 132 are described herein; however, this disclosure is not limited to these examples, as various modifications can be made to the embodiments described herein.

[0089] In some examples, one or more aspects of notebook application 132 may be implemented by notebook application 332 of server computing system 300, which can be remotely located to organize, manage, create, and interact with content in response to receiving input from a user. In some examples, one or more aspects of notebook application 332 may be implemented by notebook application 132 of computing device 100 to organize, manage, create, and interact with content in response to receiving input from a user.

[0090] According to examples of this disclosure, document application 134 can be executed by computing device 100 to provide a user of computing device 100 with a way to organize, manage, create, and interact with content, particularly content selected or chosen by the user. Document application 134 can be any kind of application related to documents (e.g., in text or visual formats) and can include word processing applications, spreadsheet applications, presentation applications, visual applications, portable document format file applications, etc. In some implementations, notebook application 132, document application 134, and document extractor application 136 can interact with each other. For example, content from a document created via document application 134 can be uploaded or stored for use by notebook application 132. In some implementations, notebook application 132 can be configured to generate documents (e.g., reports, outlines, presentations, spreadsheets) that are compatible with document application 134 (opened by or exported to document application 134). In some implementations, notebook application 132 can be configured to generate the document (e.g., report, outline, presentation, spreadsheet) based on a document template generated via document extractor application 136, giving the document a specific style, format, and / or intent.

[0091] In some examples, document application 134 may be a dedicated application specifically designed to provide a particular service. In other examples, document application 134 may be a general-purpose application (e.g., a web browser) and may enable access to a variety of different services via network 400.

[0092] According to examples of this disclosure, document extractor application 136 can be executed by computing device 100 to provide a user of computing device 100 with document templates for organizing, managing, creating, and interacting with content, particularly content selected or chosen by the user. In some implementations, document extractor application 136 may be part of document application 134 and / or notebook application 132, or it may be a standalone application. Document extractor application 136 can be configured to dynamically interact based on various user inputs. Example implementations of document extractor application 136 are described herein; however, this disclosure is not limited to these examples, as various modifications can be made to the embodiments described herein.

[0093] In some examples, one or more aspects of document extractor application 136 may be implemented by document extractor application 336 of server computing system 300, which may be remotely located to provide document templates for organizing, managing, creating, and interacting with content in response to receiving input from a user. In some examples, one or more aspects of document extractor application 336 may be implemented by document extractor application 136 of computing device 100 to provide document templates for organizing, managing, creating, and interacting with content in response to receiving input from a user.

[0094] In some example embodiments, computing device 100 includes a location determination device 140. The location determination device 140 can determine the current geographic location of computing device 100 and transmit such geographic location to server computing system 300 via network 400. The location determination device 140 can be any device or circuit for analyzing the location of computing device 100. For example, the location determination device 140 can determine the actual or relative location using a satellite navigation positioning system (e.g., GPS, Galileo, GLONASS, BeiDou), an inertial navigation system, a dead reckoning system, based on an IP address, using triangulation and / or proximity to a cell tower or WiFi hotspot, and / or other suitable techniques for determining the location of computing device 100.

[0095] The computing device 100 may include an input device 150 configured to receive input from a user, and may include one or more of the following: a keyboard (e.g., a physical keyboard, a virtual keyboard, etc.), a mouse, a joystick, buttons, switches, an electronic pen or stylus, a gesture recognition sensor (e.g., recognizing user gestures, including movements of body parts), an input sound device or a voice recognition sensor (e.g., a microphone for receiving voice input such as voice commands or voice queries), a trackball, a remote controller, a mobile phone (e.g., a cellular phone or smartphone), a tablet PC, a pedal or foot switch, a virtual reality device, etc. The input device 150 may also be embodied, for example, by a touch-sensitive display with touchscreen capability. For example, the input device 150 may be configured to receive input from a user associated with the input device 150 for selecting content to be organized or managed, for selecting queries or actions for content selected or chosen by the user, for uploading multiple training documents to generate a document template that can be applied to multiple source documents, for selecting the multiple source documents to generate an output document based on the document template, etc.

[0096] The computing device 100 may include a display device 160 that displays information viewable by a user (e.g., a user interface screen). For example, the display device 160 may be a non-touch-sensitive display or a touch-sensitive display. The display device 160 may include, for example, a liquid crystal display (LCD), a light-emitting diode (LED) display, an organic light-emitting diode (OLED) display, an active-matrix organic light-emitting diode (AMOLED), a flexible display, a 3D display, a plasma display panel (PDP), a cathode ray tube (CRT) display, etc. However, this disclosure is not limited to these example displays and may include other types of displays. The display device 160 may be used by an application system 130 located at the computing device 100 to display input-related information to a user (e.g., information related to documents, notes, projects, document templates, user interface screens with user interface elements selectable by the user, etc.).

[0097] The computing device 100 may include an output device 170 to provide output to a user, and may include one or more of the following: an audio device (e.g., one or more speakers), a haptic device (e.g., a vibration device) for providing haptic feedback to the user, a light source (e.g., one or more light sources, such as LEDs, for providing visual feedback to the user), a thermal feedback system, etc.

[0098] The computing device 100 may include a capture device 180 capable of capturing media content according to various examples of this disclosure. For example, the capture device 180 may include an image capture device 182 (e.g., a camera) configured to capture images (e.g., photographs, videos, etc.). For example, the capture device 180 may include a sound capture device 184 (e.g., a microphone) configured to capture sound or audio (e.g., audio recording). The media content captured by the capture device 180 may be transmitted, for example, via network 400 to one or more of the following: server computing system 300, content data repository 350, user data repository 360, and machine learning model data repository 370. For example, in some implementations, the media content captured by the capture device 180 may be selected by a user as source content for creating notes for a project. For example, the media content may be provided as input to one or more machine learning models to generate notes.

[0099] According to an example embodiment of this disclosure, server computing system 300 may include one or more processors 310 and one or more memory devices 320 as described herein. Server computing system 300 may also include application system 330, which is similar to application system 130 described herein.

[0100] For example, application system 330 may include: notebook application 332, which performs functions similar to those discussed herein with respect to notebook application 132; document application 334, which includes applications similar to those discussed above with respect to document application 134; and document extractor application 336, which performs functions similar to those discussed herein with respect to document extractor application 136. In some implementations, one or more machine learning models (e.g., generative machine learning models, large language models, etc.) associated with application system 330 may be configured to organize, manage, create, and interact with content based on source content selected or chosen by the user. For example, one or more machine learning models (e.g., generative machine learning models, large language models, etc.) associated with application system 330 may be configured to perform a first action (e.g., generating a summary or document guide for source content selected by the user), while computing device 100 may be configured to perform a second action (e.g., generating a suggestion action, generating an outline or study guide based on multiple notes saved to the draft area). For example, one or more machine learning models (e.g., generative machine learning models, large language models, etc.) associated with application system 130 may be configured to perform a first action (e.g., uploading source content selected by the user), while server computing system 300 may be configured to perform a second action (e.g., generating a document template based on the uploaded source content). For example, a specific action to be performed by application system 330 may vary depending on network conditions (e.g., available bandwidth, channel utilization status, latency status, throughput, etc.). In some implementations, one or more machine learning models associated with application system 330 may be configured to process user input to generate information (e.g., semantic information), which may then be provided as input to one or more other machine learning models associated with application system 330 (e.g., generative machine learning models, large language models, etc.) to generate content to be utilized for projects used in notebook application 132 and / or notebook application 332.

[0101] Examples of this disclosure also relate to computer-implemented methods for providing user interfaces for organizing, managing, and creating content by implementing one or more machine learning-based models for source content selected by the user. Figure 2 A flowchart illustrating an exemplary, non-limiting computer-implemented method according to one or more example embodiments of the present disclosure is shown. Figure 3 A block diagram of a notebook application according to one or more example embodiments of the present disclosure is shown.

[0102] Figure 2The flowchart illustrates a method 2000 for providing a user interface for organizing, managing, and creating content by implementing one or more machine learning-based models for source content selected by the user. Although shown in a specific order or sequence, the order of these processes can be modified unless otherwise specified. Therefore, the illustrated embodiments should be understood as examples only, and the illustrated processes can be executed in different orders, and some processes can be executed in parallel. Additionally, one or more processes may be omitted in various embodiments. Therefore, not all processes are necessary in every embodiment. Other process flows are possible.

[0103] refer to Figure 2 In operation 2100, method 2000 includes a computing device receiving input from a user relating to the selection of source content. As described herein, the computing device may be embodied as computing device 100, server computing system 300, or a combination thereof. For example, the input may be provided by a user via input device 150. For example, the input may be provided by selecting specific files or documents uploaded to the computing device for use by notebook application 132. In some implementations, the content may be uploaded from local storage, from another application (e.g., a portable document file application), from copied text, or from a website. These selected files, text, documents, etc., may be referred to as source content. In some implementations, the source content may be a subset of a larger content corpus. For example, the input may be provided or entered into notebook application 132 or notebook application 332.

[0104] In some implementations, the response to the input selecting the source content can be processed at computing device 100 without involving server computing system 300. In some implementations, the input selecting the source content can be transmitted from computing device 100 to server computing system 300, and at least a portion of the response to the input can be processed by server computing system 300. For example, the input related to the selection of the source content can be provided at computing device 100, and server computing system 300 can be configured to perform an operation in response to receiving an instruction from the input.

[0105] In operation 2200, the computing device can be configured to implement one or more machine learning models for selected source content to generate a document guide. In some implementations, the document guide (source guide) generated by the one or more machine learning models may include a summary of the source content and key topics related to the source content. In some implementations, the document guide may further include one or more suggested queries (e.g., questions), which may be provided as selectable user interface elements.

[0106] For example, the computing device can obtain information indicating that the user has selected source content. The computing device can process the source content using one or more machine learning models (e.g., one or more large language models) to obtain language output. The computing device can then use the one or more machine learning models (e.g., one or more large language models) to generate summary output. Specifically, the machine learning large language model can be trained to process various outputs to generate language output. For example, the machine learning large language model can process embeddings generated by a machine learning embedding generation model, portions of the source content identified using that embedding generation model, language output generated using that machine learning large language model or some other model, etc.

[0107] In operation 2300, the computing device can be configured to receive input for performing an action on the document guide. In operation 2400, the computing device can be configured to perform the action in response to receiving the input. For example, the input may be a selection of a suggested query, and the action may include providing an answer to the question by implementing one or more machine learning models for the source content. For example, the input may be text input asking a question, and the action may include providing an answer to the question by implementing one or more machine learning models for the source content. For example, the input may be a selection of a portion of the summary, and the action may include providing output indicating the specific source in the source content that was relied upon in generating the text associated with the selection of that portion of the summary.

[0108] refer to Figure 3 Notebook application 3100 (which may correspond to notebook application 132 and / or notebook application 332) may include a conditional parameter generator 3110, one or more sequence processing models 3120, one or more large language models 3130, and one or more generative machine learning models 3140. Notebook application 3100 may be as described above for... Figure 2 As discussed in operations 2100 and 2300, input 3200 is received from the user. The condition parameter generator 3110 can be configured to generate condition parameters at least in part based on the input, wherein these condition parameters provide values ​​for one or more conditions associated with the content to be generated, which is at least in part related to the input 3200 and the source content 3400 selected by the user.

[0109] For example, source content 3400 may include any kind of document (e.g., in digital form) and may include books, product manuals, legal opinions, academic papers, proprietary data files, patent documents, web pages, emails, forum posts, social media posts, videos, images, geographic information, or any other type or manner of content that can be stored or accessed digitally (e.g., in a database, storage device, etc.). In some implementations, source content 3400 may be stored in content data repository 350 by the user selecting certain documents, images, or other content to be stored in content data repository 350. In some implementations, source content 3400 may be stored at computing device 100 or server computing system 300.

[0110] To generate conditional parameters, conditional parameter generator 3110 can be configured to retrieve the values ​​of one or more conditions associated with the input. For example, to generate conditional parameters, conditional parameter generator 3110 can be configured to extract the values ​​of one or more conditions from the input. The input may include information indicating the user's intent or request. In some implementations, conditional parameter generator 3110 (or one or more sequence processing models 3120 or one or more large language models 3130) can be configured to extract information from the input 3200 to identify the values ​​of one or more conditions, and conditional parameter generator 3110 can be configured to generate conditional parameters based on the extracted values. For example, the input itself may identify the color to be used for the title in the generated document (e.g., "Title uses blue font"), or identify attributes or features that can be used to generate conditional parameters (e.g., "circular bullet points") that are used to generate a document related to the source content.

[0111] To generate conditional parameters, conditional parameter generator 3110 can be configured to infer the values ​​of one or more conditions from the input. The input may include information indicating the user's intent or request. In some implementations, conditional parameter generator 3110 (or the one or more sequence processing models 3120 or the one or more large language models 3130) can be configured to infer information from the input 3200 to identify the values ​​of the one or more conditions, and conditional parameter generator 3110 can be configured to generate conditional parameters based on the inferred values. For example, the input may include a reference to the length ("short," "long," etc.) of the summary to be generated or other documents generated based on the source content, and conditional parameter generator 3110 (or the one or more sequence processing models 3120 or the one or more large language models 3130) can be configured to infer values ​​based on the input. For example, input requesting notebook application 3100 to generate a "short" article may infer values ​​for approximately 500 words, while a "long" article may be associated with values ​​for approximately 2000 words. For example, notebook application 3100 can be configured to determine inferred values ​​based on information obtained via external content 3300.

[0112] In some implementations, the condition parameter generator 3110 may be configured to infer the value of one or more conditions from the input by providing the input to one or more sequence processing models 3120, wherein the one or more sequence processing models 3120 are configured to output the value of the one or more conditions in response to or based on the query. The one or more sequence processing models 3120 may include one or more machine learning models configured to process and analyze sequence data and data appearing in a specific order or sequence, including time series data, natural language text, or any other data having a time or sequential structure.

[0113] The one or more sequence processing models 3120 can receive input including text and segment the input by decomposing the text sequence into small units (lexical units) to provide a structured representation of the input sequence. The one or more sequence processing models 3120 can represent these lexical units as vectors in a continuous vector space by mapping each lexical unit to a high-dimensional vector, where the relationships between lexical units (words) are reflected in the geometric relationships between their corresponding vectors. For example, the one or more sequence processing models 3120 can receive input including the text “How did the Cold War end?” and segment the input by decomposing the text sequence into small units (lexical units) (e.g., “How”, “Cold War”, and “end”) to provide a structured representation of the input sequence. In word embeddings, semantically similar words are closer together in the vector space. For example, vectors for "war" and "battle" may be close to each other because of their semantic relationship, while vectors for "war" and "peace" may be far apart compared to vectors for "war" and "battle".

[0114] The one or more large language models 3130 may be, or otherwise include, models that have been trained in some way on a large language training data corpus, providing the one or more large language models 3130 with the ability to perform a variety of language tasks. For example, the one or more large language models 3130 may be trained to perform summarization tasks, dialogue tasks, simplification tasks, opposing viewpoint tasks, etc. In particular, the one or more large language models 3130 may be trained to process various outputs to generate language output. For example, the one or more large language models 3130 may process embeddings generated by a machine learning-based embedding generation model, portions of source content (e.g., document chunks) identified using the embedding generation model, language output generated using the one or more large language models 3130 or some other model, etc.

[0115] The one or more generative machine learning models 3140 may include deep neural networks or generative adversarial networks (GANs), variational autoencoders, stable diffusion machine learning models, visual Transformers, neural radiation fields (NeRF), etc., to generate content (e.g., summaries, responses to queries, etc.) with values ​​for conditions associated with one or more features. For example, the computing device may include a database (e.g., a machine learning model data repository 370) configured to store multiple generative machine learning models associated with multiple different types of content (e.g., different genres or topics, different kinds of content (including image content, video, and text), different styles of content including outlines, reports, spreadsheets, etc.). In some implementations, the computing device may be configured to retrieve from the one or more generative machine learning models 3140 a generative machine learning model associated with a specific type of content related to the input.

[0116] In some implementations, the one or more generative machine learning models 3140 can be trained on a large content dataset (e.g., a large language training data corpus) having corresponding information about conditions associated with the content. During training, the one or more generative machine learning models 3140 learn the relationships between elements in the output (e.g., content) and the conditions that influence them. This may involve the computing device adjusting the internal parameters of each generative machine learning model to generate realistic or accurate content (e.g., grammatically correct content, coherent content, etc.) based on the training data. The one or more generative machine learning models 3140 can be trained on one or more training datasets including multiple reference images of the location. The one or more training datasets may include values ​​for the one or more conditions.

[0117] In some implementations, the one or more generative machine learning models 3140 are configured to: generate a document guide 3500 in response to receiving a selection of source content 3400; and / or generate response content 3600 based on conditional parameters (and corresponding values ​​for the one or more conditions) for making a decision on generating content, the response content corresponding to content generated in response to the input for performing an action on the document guide, etc.

[0118] In some implementations, server computing system 300 may provide (transmit) the generated content or a portion thereof to computing device 100, or server computing system 300 may enable computing device 100 to access the generated content. For example, document guide 3500 may be generated at server computing system 300 and stored at one or more computing devices (e.g., computing device 100, external computing device 200, server computing system 300, external content 500, content data repository 350, user data repository 360, etc.).

[0119] In some implementations, after the documentation guide is generated and / or after an action is performed on the documentation guide, the user can provide feedback or further input related to the content generated based on the provided source content and / or via a user-provided query, and one or more of operations 2100 to 2400 can be repeated.

[0120] Examples disclosed herein also relate to user-facing aspects through which users can manage, organize, and create content via a notebook application configured to implement one or more machine learning-based models for source content selected by the user. For example, Figures 4A to 4H Example actions that can be implemented for a project according to one or more example embodiments of this disclosure are shown, in which a document guide is generated based on source content selected by a user via one or more machine learning models.

[0121] For example, Figure 4A A first user interface screen (e.g., a startup user interface screen, a startup graphical user interface, etc.) of a notebook application according to one or more example embodiments of the present disclosure is shown.

[0122] exist Figure 4A In this context, the first user interface screen 4100 depicts a user interface (e.g., a startup screen) that provides information about the notebook application 3100. Specifically, the notebook application 3100 is configured to present the first user interface screen 4100 for display, which includes various information 4110 about the features available in the notebook application 3100.

[0123] like Figure 4B As shown, the notebook application 3100 is further configured to display a second user interface screen 4200, which includes a first user interface element 4210. For example, the first user interface element 4210 is associated with enabling a user to create a new notebook (new project) through which the user can manage, organize, and create content based on source content that the user can select or curate.

[0124] like Figure 4C As shown, the notebook application 3100 is further configured to present a third user interface screen 4300 for display in response to a user providing input for creating a new notebook (e.g., via selection of a first user interface element 4210). The third user interface screen 4300 includes a first portion 4310 having a plurality of selectable user interface elements corresponding to locations from which source content can be uploaded. For example, a first user interface element 4312 corresponds to storage space that can be associated with a local computing device or a remote server system (e.g., a cloud server) or another storage device (e.g., a portable storage device). For example, a second user interface element 4314 corresponds to a portable document format file, a third user interface element 4316 corresponds to copied text, and a fourth user interface element 4318 corresponds to content that can be uploaded from a specific website or URL.

[0125] like Figure 4D As shown, notebook application 3100 is further configured to respond to requests targeting... Figure 4C The selection of one of a plurality of selectable user interface elements corresponding to a location from which source content can be uploaded is presented on a fourth user interface screen 4400 for display. The fourth user interface screen 4400 includes a first portion 4410 having a plurality of selectable source content items (e.g., multiple documents, images, videos, etc.). For example, a first user interface element 4412 corresponds to a first selected document, a second user interface element 4414 corresponds to a second selected document (e.g., a portable document format file), and a third user interface element 4416 corresponds to a third selected document. Figure 4D It shows that users can curate or select specific items from source content, which can be used to create notebooks or projects, and can be relied upon by one or more machine learning models as input data for organizing, managing, creating, and so on.

[0126] like Figure 4E As shown, notebook application 3100 is further configured to respond to requests from... Figure 4D The selection of one or more content items from the described plurality of source content items is presented on a fifth user interface screen 4500 for display. The fifth user interface screen 4500 includes a first portion 4510, a second portion 4520, a third portion 4530, and a fourth portion 4540. Each portion of the fifth user interface screen 4500 may correspond to a portion or panel of the fourth user interface screen 4400 and may be associated with different functionalities.

[0127] For example, the first section 4510 corresponds to the document guide (also known as the source guide), which includes a summary section 4512 and a key topic section 4514. The notebook application 3100 can be configured to be based on the selected source content (e.g., for...). Figure 4D As described in this article, through... Figure 3 One or more machine learning models are implemented as described to generate content (e.g., text description) associated with the summary section 4512. For example, the summary section 4512 can provide a brief summary associated with one or more content items from the content items that include the selected source content. Similarly, the notebook application 3100 can be configured to generate content based on the selected source content (e.g., as described for...). Figure 4D As described in this article, through... Figure 3 One or more machine learning models are implemented as described to generate content (e.g., text descriptions) associated with the key topic section 4514. For example, the key topic section 4514 may include one or more user interface elements that identify a topic or important subject associated with one or more content items from the content items that include the selected source content. Furthermore, the notebook application 3100 can be configured to generate output in response to the selection of one of the user interface elements in the key topic section 4514. For example, the output may be a text summary or text explanation of the key topic corresponding to the selected user interface element. This output may be provided in a separate user interface screen or in another part of a fifth user interface screen generated by the notebook application 3100 in response to the selection of one of the user interface elements in the key topic section 4514.

[0128] For example, the second part 4520 corresponds to a source content part (e.g., a context window) that includes information 4522 from at least a portion of content items in the source content. The notebook application 3100 can be configured to reproduce at least a portion of the content items from the source content in the second part 4520. In some implementations, the content in the source content part may correspond to a portion of the content items on which the summary part 4512 is relied upon.

[0129] For example, the third part 4530 corresponds to a note section (e.g., a draft area), which may include one or more notes that may be generated via various methods as described herein (e.g., automatically generated by the notebook application 3100, manually entered by the user, automatically generated by the notebook application 3100 in response to the selection of a user interface element corresponding to an action to be performed, etc.).

[0130] For example, part 4540 corresponds to a query section, which may include one or more user interface elements for submitting or providing queries to the notebook application 3100 regarding the source content. For example, part 4540 includes multiple user interface elements 4542 corresponding to suggested questions or actions related to the source content. For example, the notebook application 3100 may be configured to generate these suggested questions or actions based on information included in the source content. For example, the notebook application 3100 may be configured to additionally generate these suggested questions or actions based on conversation history (e.g., previous questions or queries), user data (e.g., user preferences, user attributes, etc.), and other contextual information. Part 4540 may further include a text input box 4544 through which a user can provide input (e.g., via keyboard, via voice input, etc.) to query the notebook application 3100. Part 4540 may further include a user interface element 4546 indicating the number of content items included in the source content. For example, in Figure 4E In the middle, user interface element 4546 indicates that the notebook application 3100 depends on three sources when generating the summary section 4512.

[0131] refer to Figure 4F The example user interface screen illustrates input questions and output responses related to the source content. For example, in Figure 4F In this context, notebook application 3100 is further configured to respond to receiving a query (e.g., via...). Figure 4E The text input field 4544 contains the text query entered, and a sixth user interface screen 4600 is displayed for this purpose. For example, the sixth user interface screen 4600 includes a first part 4610 corresponding to a dialog section, a second part 4620 corresponding to a source section, and a third part 430 corresponding to a note section (e.g., a draft area).

[0132] For example, the first part 4610 includes a prompt area 4612 corresponding to the text query and a response area 4614 corresponding to the response to the text query. In some implementations, the notebook application 3100 is configured to generate the response by implementing one or more machine learning models in response to receiving the text query as input and referring to source content 3400. For example, if a user enters a question (e.g., “How did the Cold War affect American foreign policy?”) via text input box 4544, such as for Figure 4EAs described, the notebook application 3100 can be configured to provide a sixth user interface screen 4600 and generate a response as indicated in response area 4614. As indicated in response area 4614, the number of references (source content items) that the one or more machine learning-based model depends on to generate the response can be indicated by first user interface element 4616. Figure 4F In the example, three references are used to generate the response. The response area 4614 further includes a selectable second user interface element 4618, which, when selected, causes the response to be saved as a note to a third part 4630, which corresponds to a note section (e.g., a draft area), which may include one or more notes that may be generated via various methods as described herein (e.g., automatically generated by the notebook application 3100 in response to the selection of the second user interface element 4618, etc.).

[0133] In some implementations, one or more portions of the response area may include selectable information that, when selected, allows for the display of additional information related to the selected information. For example, in Figure 4F In the text, the phrase "policy of containment" can be highlighted, bolded, underlined, or displayed in a visually distinct manner to indicate that the text is selectable (e.g., a clickable label) and that additional information related to the text is available. The notebook application 3100 can be configured to provide this additional information in response to the selection of the text (e.g., by implementing one or more machine learning models based on the selected source content) to provide additional information related to the text.

[0134] The second part 4620 may correspond to the source part and include content items 4622 containing the source content. In some implementations, each content item 4622 may correspond to a content item that the one or more machine learning models rely on when generating the response. In some implementations, the notebook application 3100 may be configured to dynamically modify or regenerate the response in the response area 4614 in response to receiving additional content items to be added as source content via user interface element 4624. Alternatively, in some implementations, the notebook application 3100 may be configured to dynamically modify or regenerate the response in the response area 4614 in response to receiving a deselection of content items from the list of content items in the second part 4620 via user interface element 4626 (e.g., by deselecting checkboxes of one or more content items in the second part 4620).

[0135] refer to Figure 4GAccording to the examples in this disclosure, the example user interface screen includes an example notes section (draft area) for the project. For example, in Figure 4G In this context, the notebook application 3100 is further configured to respond to receiving a selection of a second user interface element 4618 (e.g., as...). Figure 4F As shown), a seventh user interface screen 4700 is presented for display. When the second user interface element is selected, the response is saved as note 4712 to a third portion 4710, which corresponds to a note area (e.g., a draft area). This note area may include one or more notes, which may be generated via various methods as described herein (e.g., automatically generated by the notebook application 3100 in response to the selection of the second user interface element 4618, etc.). Figure 4G In this context, user interface element 4714 indicates the number of content items that the one or more machine learning models rely on when generating a response to note 4712. Furthermore, user interface element 4714 can be configured to be selectable, such that in response to user interface element 4714 being selected, a list of content items (references) from the source content used to generate the response can be displayed.

[0136] refer to Figure 4H According to the examples in this disclosure, the example user interface screen includes a notes section (draft area) for a project. For example, in Figure 4H In this context, the notebook application 3100 is further configured to, in response to receiving a selection of a content item 4816 from a list 4814 of content items (references), present an eighth user interface screen 4800 for display. This list of content items is derived from the source content and is generated by one or more machine learning models, stored in a note 4812, which is provided in a first portion 4810 for display. The eighth user interface screen 4800 further includes a second portion 4820 corresponding to the source portion. Figure 4H In this context, the notebook application 3100 is configured to, in response to receiving a selection of a content item 4816 from a list 4814 of various content items (references), provide information 4822 related to the selected item 4816 for display in a second part 4820. The list of content items is derived from the source content and is generated by one or more machine learning models and stored in the notebook 4812.

[0137] In some implementations, information 4822 may include information from the content item 4816 used to generate the response. For example, notebook application 3100 may be configured to reference metadata associated with the response to reference information 4822. This metadata may indicate the location of information from a content item used to generate the response. Furthermore, information 4822 may correspond to or include specific text from the content item that was relied upon in generating the response. For example, notebook application 3100 may be configured to display this specific text in a visually distinct manner (e.g., highlighted, bold, enlarged font size, underlined, italicized, etc.) in the second part 4820. For example, notebook application 3100 may be configured to display additional text appearing before and / or after this specific text in the second part 4820. This additional information can provide the user with further context regarding the information relied upon in generating the response. For example, notebook application 3100 can be configured to mark specific content items that are relied upon when generating the response in note 4812, and to mark specific text segments within those specific content items that are relied upon when generating the response in note 4812. Therefore, users can easily and intuitively identify where support for the response can be found within the content items.

[0138] In some implementations, information 4822 from the selected content item 4816 used to generate the response can be truncated or displayed in its entirety. For example, when information 4822 is less than a threshold, the entire text from the selected content item 4816 can be displayed in a second part 4820 and can be used by the one or more machine learning models to generate a response (e.g., a response to a text query). For example, when information 4822 is greater than the threshold, notebook application 3100 can be configured to implement a semantic retrieval method to determine specific text segments from the entire selected content item 4816 that are relevant to the user query (e.g., a text query). In this example, these relevant text segments (rather than the entire information from the content item) are relied upon by the one or more machine learning models to generate a response to the user query (e.g., the text query).

[0139] The examples disclosed herein relate to additional user-facing aspects through which users can manage, organize, and create content via a notebook application configured to implement one or more machine learning-based models for source content selected by the user. For example, Figures 5A to 5B Example actions that can be implemented for a project according to one or more example embodiments of this disclosure are shown, in which notes are generated based on source content selected by a user via one or more machine learning models.

[0140] For example, Figure 5AA first user interface screen of a notebook application according to one or more example embodiments of the present disclosure is shown. For example, in Figure 5A In this context, the first user interface screen 5100 includes a first portion 5110, a second portion 5120, and a third portion 5130. The first portion 5110 corresponds to a note portion (e.g., a draft area), which may include one or more notes 5112, which may be generated via various methods as described herein (e.g., automatically generated by the notebook application 3100, manually entered by the user, automatically generated by the notebook application 3100 in response to the selection of a user interface element corresponding to an action to be performed, etc.).

[0141] The second part 5120 corresponds to the source content part (e.g., a source guide or context window), which may include one or more sources 5122 (e.g., content items of source content 3400 that the one or more machine learning models rely on when generating information included in the one or more notes 5112). In some implementations, the notebook application 3100 may be configured to generate notes based on selections of at least a portion of information from a content item provided in the second part 5120, which are then saved as notes to the first part 5110. For example, Figure 5A The selected text 5124 (e.g., highlighted text) that has been chosen by the user is displayed.

[0142] For example, the third part 5130 corresponds to a query section, which may include one or more user interface elements for submitting or providing queries to the notebook application 3100 regarding the source content. For example, the third part 5130 includes multiple user interface elements 5132 corresponding to suggested questions or actions related to the source content. For example, the notebook application 3100 may be configured to generate these suggested questions or actions based on information included in the source content and / or based on information displayed in the second part 5120. The third part 5130 may further include a text input box 5134 through which a user can provide input (e.g., via keyboard, via voice input, etc.) to query the notebook application 3100. The third part 5130 may further include a user interface element 5136 indicating the number of content items included in the source content. For example, in Figure 5A In the text, user interface element 5136 indicates that notebook application 3100 relied on three sources when generating the one or more notes 5112.

[0143] In some implementations, the plurality of user interface elements 5132 can be configured to change dynamically based on actions performed on the first user interface screen 5100. For example, the notebook application 3100 can be configured to dynamically change, modify, delete, or add user interface elements in the third part 5130 based on actions performed on the source content (e.g., on content items provided in the second part 5120 for display). Figure 5A In this context, the notebook application 3100 can be configured to dynamically change user interface elements in the third part 5130 based on (in response to) selections of text (e.g., selected text 5124) from one or more sources 5122. For example, as... Figure 5A As indicated, these actions may include summarizing the selected text into notes, adding citations to the notes, requesting additional information about the selected text 5124, or suggesting related ideas. For example, notebook application 3100 may be configured to generate notes summarizing the selected text in response to receiving a selection of user interface element 5132a corresponding to the action of summarizing the selected text into notes. For example, notebook application 3100 may be configured to add content corresponding to the selected text to an existing note in response to receiving a selection of user interface element 5132b corresponding to the action of adding citations to the notes.

[0144] For example, Figure 5B A second user interface screen of a notebook application according to one or more example embodiments of the present disclosure is shown. For example, in Figure 5B In the second user interface screen 5200, there are a first part 5210, a second part 5220, and a third part 5230, each of which can correspond to... Figure 5A Part 1 5110, Part 2 5120 and Part 3 5130.

[0145] For example, regarding Figure 5A As described, the notebook application 3100 can be configured to generate notes summarizing the selected text in response to receiving a selection of a user interface element 5132a corresponding to the action of summarizing the selected text into notes. Figure 5B The generated notes 5214, which have been saved to a first part 5210, are shown. This first part includes one or more notes 5212. Furthermore, in some implementations, after the generated notes 5214 are saved to the first part 5210, data from... Figure 5A The multiple user interface elements 5132 can be configured to dynamically change back to their previous state. Figure 5B The multiple user interface elements 5232 shown.

[0146] The examples disclosed herein relate to additional user-facing aspects through which users can manage, organize, and create content via a notebook application configured to implement one or more machine learning-based models for source content selected by the user. For example, Figures 6A to 6B Example actions that can be implemented for a project according to one or more example embodiments of this disclosure are shown, in which notes are generated based on source content selected by a user via one or more machine learning models.

[0147] For example, Figure 6A A portion of a first user interface screen of a notebook application according to one or more example embodiments of the present disclosure is shown. For example, in Figure 6A The diagram illustrates a first portion 6110 and a second portion 6120 of a user interface screen. The first portion 6110 corresponds to a notes section (e.g., a draft area), which may include multiple notes that may have been generated via various methods as described herein (e.g., automatically generated by the notebook application 3100, manually entered by the user, automatically generated by the notebook application 3100 in response to a selection of a user interface element corresponding to an action to be performed, etc.). For example, the first portion 6110 may indicate how a particular note was created (e.g., as a saved response, as a written note written by the user, as a document generated from other notes, etc.).

[0148] For example, the second part 6120 corresponds to a query section, which may include one or more user interface elements for submitting or providing queries to the notebook application 3100 regarding the source content or the multiple notes. For example, the second part 6120 includes multiple user interface elements 6122 corresponding to suggested questions or actions related to the source content or the multiple notes. For example, the notebook application 3100 may be configured to generate these suggested questions or actions based on information included in the source content and / or based on information displayed in the first part 6110. The second part 6120 may further include: a text input box through which a user can provide input (e.g., via keyboard, via voice input, etc.) to query the notebook application 3100; and user interface elements indicating the number of content items including the source content, etc.

[0149] exist Figure 6A In the example, notebook application 3100 can be configured to dynamically change user interface elements in second section 6120 based on (in response to) a selection of one or more notes 6112 from the plurality of notes provided in first section 6110. For example, as Figure 6AAs indicated, one or more notes can be selected via user input (e.g., via drag-and-drop input, via selecting checkboxes, etc.), and in response to the selection of the one or more notes, actions can include creating content based on the selected notes 6114 (e.g., creating a study guide, creating an outline, creating a spreadsheet, creating a presentation, etc.), suggesting related ideas, etc. For example, notebook application 3100 can be configured to generate notes corresponding to the outline of the content from the selected notes 6114 in response to receiving a selection of a user interface element 6122a corresponding to the action of creating an outline and saving the outline as a note. For example, notebook application 3100 can be configured to implement one or more machine learning models to generate notes corresponding to the selected notes 6114 in response to receiving a selection of a user interface element corresponding to the action of creating content for the one or more selected notes and saving the content as a note.

[0150] For example, Figure 6B A portion of a second user interface screen for a notebook application according to one or more example embodiments of this disclosure is shown. For example, in Figure 6B In the middle, the first part 6210 can correspond to Figure 6A Part 1, 6110.

[0151] For example, regarding Figure 6A As described, the notebook application 3100 can be configured to generate notes based on one or more selected notes by implementing one or more machine learning models, wherein the selected notes may correspond to source content (e.g., source content selected by the user and used as input to generate the note). For example, the generated notes may be based on previously selected notes. Figure 6A The selected notes are summarized or outlined as described. Notebook application 3100 can be configured to generate the generated notes 6214 based on the selected notes 6114 by implementing one or more machine learning models and in response to receiving a selection of a user interface element (e.g., user interface element 6122a) corresponding to the action of summarizing the selected notes 6114 into the generated notes 6214, wherein the selected notes 6114 may correspond to source content (e.g., source content selected by the user and used as input for generating the notes). Figure 6B The generated notes 6214, which have been saved to a first portion 6210, are shown. This first portion includes one or more other notes 6212. Furthermore, in some implementations, after the generated notes 6214 are saved to the first portion 6210, data from... Figure 6A The multiple user interface elements 6122 can be configured to dynamically change back to their previous states.

[0152] In some implementations, the notebook application 3100 can be configured to send a document to another application (e.g., a word processing application, presentation application, spreadsheet application, social media application, etc.) via selection of user interface elements, enabling the generated notes 6214 to be exported to other applications. In some implementations, the notebook application 3100 can be configured to share a document and / or source content with another user via selection of user interface elements, enabling the generated notes 6214 and / or content items (e.g., source content 3400) to be shared with other users.

[0153] According to examples of this disclosure, notebook application 3100 can be configured to generate outputs (e.g., outlines, reports, summaries, etc.) based on source content provided to the notebook application (e.g., by a user) via one or more machine learning models. Notebook application 3100 can be configured to allow users to create various projects to complete various tasks. Each project can be configured to function like a folder, allowing users to store various information into each project. In some implementations, a separate draft area can correspond to or be dedicated to a specific project. In some implementations, notebook application 3100 can be configured to receive the source content specified by the user. Notebook application 3100 can be configured to add, delete, or modify projects based on input received from the user. Each project can be provided with a default name, a name provided by the user, or a name generated by notebook application 3100 (e.g., via one or more machine learning models) based on information stored in that project (e.g., based on the source content).

[0154] The examples disclosed herein relate to additional user-facing aspects through which users can manage, organize, and create content via a notebook application configured to implement one or more machine learning-based models for source content selected by the user. For example, Figure 7 Examples of notebooks or projects are shown, which may be represented in a specific way so that users can easily understand the content contained in the notebook or project.

[0155] In some implementations, in response to source content being provided to notebook application 3100, notebook application 3100 may be configured to automatically generate (e.g., using one or more machine learning models, one or more generative machine learning models, semantic retrieval techniques, etc.) a graphical image (e.g., an emoji, icon, etc.) or graphical animation corresponding to or representing the source content. In some implementations, the graphical image or graphical animation may be overlaid on a folder provided as a user interface element, which, when selected, causes the folder to open and display its contents to the user. Alternatively, in some implementations, in response to source content being provided to notebook application 3100, notebook application 3100 may be configured to automatically generate (e.g., using one or more machine learning models, one or more generative machine learning models, semantic retrieval techniques, etc.) a text description (name) corresponding to or representing the source content. This text description may be overlaid on a folder provided as a user interface element, which, when selected, causes the folder to open and display its contents to the user.

[0156] refer to Figure 7The notebook application 3100 may have user-specific sections 7100, which store various items in specific folders. For example, a first folder 7110 (e.g., the default folder) may be represented by a default image 7112 and have a generic name 7114 (e.g., "Default Notebook"). A second folder 7120 may be represented by a graphic image 7122 and have a text description 7124 (e.g., "Benefits") generated by machine learning and representing or corresponding to content included in the second folder 7120. For example, in response to source content being provided to the notebook application 3100, the notebook application 3100 may be configured to automatically generate (e.g., using one or more machine learning models, one or more generative machine learning models, semantic retrieval techniques, etc.) a graphic image 7122, which may correspond to or represent the source content as an emoji, icon, etc. In some implementations, a graphic image 7122 may be overlaid on a second folder 7120, which is provided as a user interface element. When selected, this user interface element causes the second folder 7120 to open and display its contents to the user. Alternatively, in some implementations, in response to the source content being provided to a notebook application 3100, the notebook application 3100 may be configured to automatically generate (e.g., using one or more machine learning models, one or more generative machine learning models, semantic retrieval techniques, etc.) a text description 7124 (name) corresponding to or representing the source content. The text description 7124 may be overlaid on the second folder 7120, which is provided as a user interface element. When selected, this user interface element causes the second folder 7120 to open and display its contents to the user.

[0157] Examples of this disclosure relate to methods for generating document templates and for providing computer-implemented user interfaces for organizing, managing, and creating content by implementing one or more machine learning-based models for source content selected by a user and the generated document template. Figure 8A A flowchart illustrating an exemplary, non-limiting computer-implemented method according to one or more example embodiments of the present disclosure is shown. Figure 8B Another flowchart of an exemplary, non-limiting computer-implemented method according to one or more example embodiments of the present disclosure is shown. Figure 9 A block diagram of a document extractor application according to one or more example embodiments of the present disclosure is shown.

[0158] Figure 8AThe flowchart illustrates a method 8000 for generating a document template, which can be used to organize, manage, and create content by implementing one or more machine learning models against training content (e.g., sample documents, training documents, etc.) selected by a user. Although shown in a specific order or sequence, the order of these processes can be modified unless otherwise specified. Therefore, the illustrated embodiments should be understood as examples only, and the illustrated processes can be performed in different orders, and some processes can be performed in parallel. Furthermore, one or more processes may be omitted in various embodiments. Therefore, not all processes are required in every embodiment. Other process flows are possible.

[0159] refer to Figure 8A In operation 8100, method 8000 includes a computing device receiving input from a user relating to the selection of sample documents (e.g., training content, training documents, etc.). As described herein, the computing device may be embodied as computing device 100, server computing system 300, or a combination thereof. For example, the input may be provided by a user via input device 150. For example, the input may be provided by selecting specific files or documents uploaded to the computing device for use by document extractor application 136. In some implementations, the content may be uploaded from local storage, from another application (e.g., a portable document file application), from copied text, or from a website. These selected files, text, documents, etc., may be referred to as training content, training documents, sample documents, etc. In some implementations, the training content may be a subset of a larger content corpus. For example, the input may be provided or input to document extractor application 136 or document extractor application 336.

[0160] In some implementations, the response to the input selecting the source content can be processed at computing device 100 without involving server computing system 300. In some implementations, the input selecting the training content can be transmitted from computing device 100 to server computing system 300, and at least a portion of the response to the input can be processed by server computing system 300. For example, the input related to the selection of the training content can be provided at computing device 100, and server computing system 300 can be configured to perform an operation in response to receiving an instruction from the input.

[0161] In operation 8200, method 8000 includes the computing device receiving input from a user requesting the selection of a generated document template regarding a sample document (e.g., training content, training document, etc.). For example, this input may be provided by the user via input device 150. In some implementations, the computing device (e.g., document extractor application 136) may be configured to provide a graphical user interface for display on a display device, through which the user can request the generation of the document template. For example, the input may be provided by selecting a user interface element associated with generating the document template. For example, the input may be provided or entered into document extractor application 136 or document extractor application 336.

[0162] In operation 8300, method 8000 includes the computing device implementing one or more machine learning models for selected training content (sample documents, training documents, etc.) to generate a document template. In some implementations, the document template generated by the one or more machine learning models may include multiple sections, multiple headings indicating different sections of the document, etc. In some implementations, the document template may be further associated with a specific style, intent, and / or format that can be inferred or extracted from the content of the training content.

[0163] For example, the computing device can obtain information indicating that the user has selected the training content. The computing device can process the training content using one or more machine learning models (e.g., one or more large language models) to obtain language output. The computing device can then use the one or more machine learning models (e.g., one or more large language models, one or more generative machine learning models, etc.) to generate summary output. Specifically, the large language model can be trained to process various outputs to generate language output. For example, the large language model can process embeddings generated by a machine learning embedding generation model, portions of the training content identified using that embedding generation model, language output generated using that large language model or some other model, etc.

[0164] In some implementations, the one or more machine learning models can be configured to determine (learn) the intent, style, and / or format of the training content, for example, via various natural language processing operations. For instance, the training content can be decomposed into lemmas (e.g., words, phrases, individual characters, etc.) and converted into embeddings (e.g., numerical vector representations) that can capture semantic information about the training content. In some implementations, each of the multiple training documents can be classified into a specific type of document (e.g., resume, PRD, outline, legal opinion, etc.). The training document can be classified based on the aggregation of lemma embeddings to create a representation of the entire training document (e.g., via embedding averaging, TF-IDF weighting, etc.). The one or more machine learning models can be configured to analyze the vocabulary in the training document (e.g., based on the frequency of certain words, the presence of specific terms, the use of domain-specific jargon, etc.). The one or more machine learning models can also be configured to determine the syntax of the training document (e.g., based on sentence structure, sentence length, use of grammatical structures, etc.), which can provide information about the document's specific style and / or intent. The one or more machine learning models can be configured to analyze the training content to determine semantic information based on the meaning of the content (e.g., the meaning of specific sentences and paragraphs in the document) to identify the specific style and / or intent of the training document. Furthermore, the one or more machine learning models can be configured to determine the context of each word relative to the entire training document.

[0165] To determine the format of the training content, one or more machine learning models can be configured to analyze the sequential structure of the training content to identify recurring patterns (e.g., identifying headings, subheadings, paragraphs, bullet points, numbered lists, and other common formatting elements). The one or more machine learning models can also be configured to learn document structure based on consistent patterns or layouts (e.g., tables, images, captions, etc.) to understand spatial relationships between different elements. The one or more machine learning models can also be configured to recognize specific formatting conventions (e.g., the use of indentation, font style, font size, etc.) to identify the document structure and perform pattern matching. The one or more machine learning models can be configured to learn and recognize specific document formats so that when a user uploads multiple training documents, a document template conforming to the document structure of those training documents can be generated. In some implementations, the training content can share common features (e.g., common formatting, common style, common intent, etc.).

[0166] In some implementations, the one or more machine learning models can be configured to receive information from the user that identifies information about the training content (e.g., labeled data, such as identifiers of document type, document style, document intent, document format, etc.). The one or more machine learning models can be trained and / or refined based on the labeled data and feedback provided by the user. The document template generated by the one or more machine learning models (e.g., one or more large language models, one or more generative machine learning models, etc.) in operation 8300 can be stored in the computing device and can be output to be presented to the user on the display device 160.

[0167] Figure 8B The flowchart shows the process for using based on Figure 8A The flowchart generates the document template, and the method 8400 generates the output document by implementing one or more machine learning models for source content (e.g., a source document) selected by the user. Although shown in a specific order or sequence, the order of these processes can be modified unless otherwise specified. Therefore, the illustrated embodiments should be understood as examples only, and the illustrated processes can be performed in different orders, and some processes can be performed in parallel. In addition, one or more processes may be omitted in various embodiments. Therefore, not all processes are required in every embodiment. Other process flows are possible.

[0168] refer to Figure 8B In operation 8500, method 8400 includes a computing device receiving input from a user relating to the selection of a source document (e.g., a source document). As described herein, the computing device may be embodied as computing device 100, server computing system 300, or a combination thereof. For example, the input may be provided by a user via input device 150. For example, the input may be provided by selecting specific files or documents uploaded to the computing device for use by document extractor application 136. In some implementations, the content may be uploaded from local storage, from another application (e.g., a portable document file application), from copied text, or from a website. These selected files, text, documents, etc., may be referred to as source content, source documents, etc. In some implementations, the source content may be a subset of a larger content corpus. For example, the input may be provided or input to document extractor application 136 or document extractor application 336.

[0169] In some implementations, the response to the input selecting the source content can be processed at computing device 100 without involving server computing system 300. In some implementations, the input selecting the source content can be transmitted from computing device 100 to server computing system 300, and at least a portion of the response to the input can be processed by server computing system 300. For example, the input related to the selection of the source content can be provided at computing device 100, and server computing system 300 can be configured to perform an operation (e.g., the generation of the output document) in response to receiving an indication of the input.

[0170] In operation 8600, method 8400 includes the computing device receiving input from a user requesting selection of source documents (e.g., source content) and a specific document template, also selectable via user input, to generate an output document. For example, this input may be provided by the user via input device 150. In some implementations, the computing device (e.g., document extractor application 136) may be configured to provide a graphical user interface for presentation on a display device, through which the user can request the generation of the output document in combination with a specific document template, also selectable via user input. For example, this input may be provided by selecting user interface elements associated with selecting the desired document template and generating the output document. In some implementations, input for generating the output document may be provided, and document extractor application 136 or document extractor application 336 may be configured to determine applicable document templates that can be applied to these source documents to generate the output document. For example, this input may be provided or input to document extractor application 136 or document extractor application 336.

[0171] In operation 8700, method 8400 includes the computing device implementing one or more machine learning models (e.g., one or more large language models, one or more generative machine learning models, etc.) for a selected source document (source content) and an identified document template to generate the output document. In some implementations, the output document generated by the one or more machine learning models may include multiple sections, multiple headings indicating different sections of the document, etc., which conform to the identified or selected document template. In some implementations, the output document may be further associated with a specific style, intent, and / or format based on the style, intent, and / or format of the document template.

[0172] For example, the computing device can obtain information indicating that the user has selected these source documents. The computing device can process these source documents using one or more machine learning models (e.g., one or more large language models) to obtain language output. The computing device can then use the one or more machine learning models (e.g., one or more large language models, one or more generative machine learning models, etc.) to generate summary output. Specifically, the large language model can be trained to process various outputs to generate language output. For example, the large language model can process embeddings generated by a machine learning embedding generation model, portions of source documents identified by the embedding generation model, language output generated using the large language model or some other model, etc.

[0173] In some implementations, the one or more machine learning-based models can be configured to determine (learn) the intent, style, and / or format of source documents, for example, via various natural language processing operations. For instance, these source documents can be decomposed into lemmas (e.g., words, phrases, individual characters, etc.) and converted into embeddings (e.g., numerical vector representations) that can capture semantic information about these source documents. In some implementations, each of the multiple source documents can be classified into a specific type of document (e.g., resume, PRD, outline, legal opinion, etc.). The source document can be classified based on the aggregation of lemma embeddings to create a representation of the entire training document (e.g., via embedding averaging, TF-IDF weighting, etc.). The one or more machine learning-based models can be configured to analyze the vocabulary in these source documents (e.g., based on the frequency of certain words, the presence of specific terms, the use of domain-specific jargon, etc.). The one or more machine learning-based models can also be configured to determine the syntax of the source documents (e.g., based on sentence structure, sentence length, use of grammatical structures, etc.), which can provide information about the specific style and / or intent of the source document. The one or more machine learning models can be configured to analyze these source documents to determine semantic information based on the meaning of the content (e.g., the meaning of specific sentences and paragraphs in the document) to identify the specific style and / or intent of the source document, etc. Furthermore, the one or more machine learning models can be configured to determine the context of each word relative to the entire source document.

[0174] The one or more machine learning models (e.g., one or more large language models, one or more generative machine learning models, etc.) can be configured to apply the document template to the multiple source documents to generate an output document. For example, the one or more machine learning models can be configured to extract first content (e.g., background information, title information, body information, conclusion information, etc.) from the multiple source documents and associate the first content with a first part of the output document (e.g., background section, title section, body section, conclusion section, etc.). Similarly, the one or more machine learning models can be configured to extract second content from the multiple source documents and associate the second content with a second part of the output document, and so on. The one or more machine learning models (e.g., one or more large language models, one or more generative machine learning models, etc.) can be configured to associate content from these source documents with specific parts of the output document based on determined semantic information, contextual information, intent information, etc., associated with the content. For example, if one or more machine learning models (e.g., based on confidence levels) determine that a portion of a source document is associated with background information about a topic or subject, then the one or more machine learning models can be configured to implement some or all of that portion in the background section of the output document.

[0175] For example, the one or more machine learning models can be configured to apply the style, intent, and / or format of the document template to content extracted from the multiple source documents. For instance, if the document template is associated with a persuasive intent and a style that clearly defines an opinion, the one or more machine learning models can be configured to generate an output document with such characteristics based on content from the multiple source documents, wherein the output document may have a document structure defined by or associated with the document template.

[0176] As another example, the one or more machine learning models can be configured to identify a document type associated with each of the multiple source documents based on the content of each of the multiple source documents. The one or more machine learning models can be configured to apply the style, intent, and / or formatting of a document template associated with the identified document type. For example, if the one or more machine learning models determine that the document type associated with these source documents is a PRD, then the one or more machine learning models can be configured to apply the style, intent, and / or formatting of a document template associated with the PRD document type to the content from the multiple source documents.

[0177] In some implementations, the one or more machine learning models can be configured to receive information from the user that identifies information about the output document and / or these source documents (e.g., labeled data, such as identifiers of document type, document style, document intent, document format, etc.). The one or more machine learning models used to generate the output document can be trained and / or refined based on the labeled data and feedback provided by the user. For example, the output document generated in operation 8700 can be stored in the computing device and can be output to be presented to the user on display device 160.

[0178] refer to Figure 9 Document extractor application 9100 (which may correspond to document extractor application 136 and / or document extractor application 336) may include a conditional parameter generator 9110, one or more sequence processing models 9120, one or more large language models 9130, and one or more generative machine learning models 9140. Document extractor application 9100 may be as described above for... Figure 8A and Figure 8B As discussed in operations 8100, 8200, 8500, and 8600, input 9200 is received from the user. The condition parameter generator 9110 can be configured to generate condition parameters at least in part based on the input, wherein these condition parameters provide values ​​for one or more conditions associated with content to be generated, which is at least in part related to the input 9200 and training content 9400 and / or source content 9500 selected by the user.

[0179] For example, training content 9400 and source content 9500 can include any kind of document (e.g., in digital form) and can include books, product manuals, legal opinions, academic papers, proprietary data files, patent documents, web pages, emails, forum posts, social media posts, videos, images, geographic information, or any other type or manner of content that can be stored or accessed digitally (e.g., in a database, storage device, etc.). In some implementations, training content 9400 and source content 9500 can be stored in content data repository 350 by the user selecting certain documents, images, or other content to be stored in content data repository 350. In some implementations, training content 9400 and source content 9500 can be stored at computing device 100 and / or server computing system 300.

[0180] To generate conditional parameters, conditional parameter generator 9110 can be configured to retrieve the values ​​of one or more conditions associated with the input. For example, to generate conditional parameters, conditional parameter generator 9110 can be configured to extract the values ​​of one or more conditions from the input. The input may include information indicating the user's intent or request. In some implementations, conditional parameter generator 9110 (or the one or more sequence processing models 9120 or the one or more large language models 9130) can be configured to extract information from the input 9200 to identify the values ​​of one or more conditions, and conditional parameter generator 9110 can be configured to generate conditional parameters based on the extracted values. For example, the input itself may identify the color to be used for the title in the generated document (e.g., "The title uses blue font"), or identify attributes or features that can be used to generate conditional parameters (e.g., "circular bullet points"), which are used to generate document templates associated with training content 9400 or output documents associated with source content 9500.

[0181] To generate conditional parameters, conditional parameter generator 9110 can be configured to infer the value of one or more conditions from the input. The input may include information indicating the user's intent or request. In some implementations, conditional parameter generator 9110 (or one or more sequence processing models 9120 or one or more large language models 9130) can be configured to infer information from the input 9200 to identify the value of one or more conditions, and conditional parameter generator 9110 can be configured to generate conditional parameters based on the inferred values. For example, the input may include a reference to the length ("short," "long," etc.) of the output document to be generated based on source content 9500, and conditional parameter generator 9110 (or one or more sequence processing models 9120 or one or more large language models 9130) can be configured to infer values ​​based on the input. For example, input to request document extractor application 9100 to generate a "standard" resume may infer approximately one page of values, while a "long" resume may be associated with approximately two to three pages of values. For example, a document extractor application 9100 can be configured to determine inferred values ​​based on information obtained via external content 9300 (e.g., a website describing the length of a resume).

[0182] In some implementations, the condition parameter generator 9110 may be configured to infer the value of one or more conditions from the input by providing the input to one or more sequence processing models 9120, wherein the one or more sequence processing models 9120 are configured to output the value of the one or more conditions in response to or based on the query. The one or more sequence processing models 9120 may include one or more machine learning models configured to process and analyze sequence data and data appearing in a specific order or sequence, including time series data, natural language text, or any other data having a time or sequential structure.

[0183] The one or more sequence processing models 9120 can receive input including text and segment the input by decomposing the text sequence into small units (lexical units) to provide a structured representation of the input sequence. The one or more sequence processing models 9120 can represent these lexical units as vectors in a continuous vector space by mapping each lexical unit to a high-dimensional vector, where the relationships between lexical units (words) are reflected in the geometric relationships between their corresponding vectors. For example, the one or more sequence processing models 9120 can receive input including the text "the bustling marketplace" extracted from training content 9400 and / or source content 9500 and segment the input by decomposing the text sequence into small units (lexical units) (e.g., "the", "bustling", and "marketplace") to provide a structured representation of the input sequence. In word embeddings, semantically similar words are closer together in the vector space. For example, vectors for "bustling" and "busy" may be close to each other because of their semantic relationship, while vectors for "bustling" and "stagnant" may be far apart compared to vectors for "bustling" and "stagnant".

[0184] The one or more large language models 9130 may be, or otherwise include, models that have been trained in some way on a large language training data corpus, providing the one or more large language models 9130 with the ability to perform a variety of language tasks. For example, the one or more large language models 9130 may be trained to perform summarization tasks, dialogue tasks, simplification tasks, opposing viewpoint tasks, etc. In particular, the one or more large language models 9130 may be trained to process various outputs to generate language output. For example, the one or more large language models 9130 may process embeddings generated by a machine learning-based embedding generation model, portions of source content or training content identified by the embedding generation model (e.g., document chunks), language output generated using the one or more large language models 9130 or some other model, etc.

[0185] The one or more generative machine learning models 9140 may include deep neural networks or generative adversarial networks (GANs), variational autoencoders, stable diffusion machine learning models, visual Transformers, neural radiation fields (NeRF), etc., to generate content (e.g., resumes, outlines, PRDs, etc.) with values ​​for conditions associated with one or more features. For example, the computing device may include a database (e.g., a machine learning model data repository 370) configured to store multiple generative machine learning models associated with multiple different types of content or multiple different types of documents (e.g., different genres or topics, different kinds of content (including image content, video, and text), different types of content (including outlines, reports, spreadsheets, resumes, PRDs, etc.)).

[0186] In some implementations, the computing device may be configured to retrieve generative machine learning models associated with a specific type of content (document) and / or document template related to the input from the one or more generative machine learning models 9140 for generating the output document.

[0187] In some implementations, the one or more generative machine learning models 9140 can be trained on a large content dataset (e.g., a large language training data corpus) having corresponding information about conditions associated with the content. During training, the one or more generative machine learning models 9140 can be configured to learn the relationships between elements in the output (e.g., content) and the conditions that influence them. This may involve the computing device adjusting the internal parameters of each generative machine learning model to generate realistic or accurate content (e.g., grammatically correct content, coherent content, etc.) based on the training data. The one or more generative machine learning models 9140 can be trained on one or more training datasets including multiple reference document templates. The one or more generative machine learning models 9140 can be trained on one or more training datasets including multiple reference output documents associated with one or more document templates. The one or more training datasets may include values ​​for the one or more conditions.

[0188] In some implementations, the one or more generative machine learning models 9140 are configured to generate a document template 9700 in response to receiving a selection of training content 9400. For example, the document template 9700 may be generated based on conditional parameters (and corresponding values ​​for the one or more conditions) to make decisions about the content used to generate the document template 9700. In some implementations, the one or more generative machine learning models 9140 are configured to generate an output document 9800 in response to receiving a selection of source content 9500 and based on a specific document template 9700, which may be selected by a user or automatically determined based on the source content 9500. For example, the output document 9800 may be generated based on conditional parameters (and corresponding values ​​for the one or more conditions) to make decisions about the content used to generate the output document 9800.

[0189] In some implementations, server computing system 300 may provide (transmit) the generated content or a portion thereof to computing device 100, or server computing system 300 may enable computing device 100 to access the generated content. For example, document template 9700 and / or output document 9800 may be generated at server computing system 300 and stored at one or more computing devices (e.g., computing device 100, external computing device 200, server computing system 300, external content 500, content data repository 350, user data repository 360, etc.).

[0190] In some implementations, after document template 9700 is generated, the user can provide feedback or further input regarding the document template 9700 generated based on the provided training content 9400 (and / or based on a query provided by the user), and one or more of operations 8100 to 8300 can be repeated. In some implementations, after output document 9800 is generated, the user can provide feedback or further input regarding the output document 9800 generated based on the provided source content 9500 (and / or based on a query provided by the user), and one or more of operations 8500 to 8700 can be repeated.

[0191] Examples disclosed herein also relate to user-facing aspects through which users can manage, organize, and create content via notebook applications and / or document extractor applications, each configured to implement one or more machine learning-based models for content selected by the user. For example, Figure 10A Figures 10F illustrate example actions that can be implemented for a project according to one or more example embodiments of the present disclosure, in which a document template is generated via one or more machine learning models based on training content selected by the user, and an output document is generated via one or more machine learning models based on source content selected by the user.

[0192] For example, Figure 10A A first user interface screen (e.g., a template builder user interface screen) of a document extractor application (which may be incorporated as part of a notebook application) according to one or more example embodiments of the present disclosure is shown.

[0193] exist Figure 10A In this context, the first user interface screen 1010 depicts a user interface (e.g., a template builder user interface screen) that provides information about the document extractor application 9100. Specifically, the document extractor application 9100 is configured to present the first user interface screen 1010 for display, which includes a visual depiction 1012 of how to create and build document templates based on training content and information 1014 about the characteristics of the document extractor application 9100.

[0194] like Figure 10A As shown, the first user interface screen 1010 further includes multiple user interface elements that can be selected (or interacted with) by a user to generate a document template. For example, a first portion of the first user interface screen 1010 includes multiple first user interface elements 1016 configured to be selected as training content for creating a document template. Figure 10AIn this context, the training content that users can choose to generate document templates includes PRD files (“Product PRD”), notes from meetings (“Brainstorming Meetings”), and notes on strategies (“Notebook LLM Strategy”).

[0195] For example, a second portion of the first user interface screen 1010 includes a second user interface element 1018, which is configured to enable a user to upload one or more training documents for generating document templates. Figure 10A As shown, the third user interface element 1019 is configured to receive input associated with creating (generating) a document template based on training content (e.g., training documents) identified (selected) by the user via the first user interface screen 1010. For example, a document extractor application 9100 may be configured to generate a document template based on training content selected by the user and in response to user input (e.g., selection of the third user interface element 1019), according to the example described herein.

[0196] Figure 10B A second user interface screen (e.g., a document template customization user interface screen) of a document extractor application (which may be incorporated as part of a notebook application) according to one or more example embodiments of the present disclosure is shown.

[0197] exist Figure 10B In this context, the second user interface screen 1020 depicts a user interface (e.g., a template builder user interface screen) associated with one or more features that enable the user to customize a document template generated by the document extractor application 9100. Specifically, the document extractor application 9100 is configured to display the second user interface screen 1020, which includes various sections and user interface elements through which the user can modify or customize the document template generated by the document extractor application 9100.

[0198] For example, the second user interface screen 1020 includes a first portion 1021 that identifies the name of the generated document template (“PRD template”).

[0199] For example, the second user interface screen 1020 includes a second section 1022 associated with a training document selected for generating the document template. For example, training documents 1023 include a "Product PRD" training document and a "Brainstorming Meeting" document. The first user interface element 1024 can be configured to allow the user to add training documents, enabling the document template to be regenerated based on the added training documents.

[0200] For example, the second user interface screen 1020 includes a third portion 1025 associated with the style of the generated document template. For example, the third portion 1025 of the second user interface screen 1020 includes a plurality of first user interface elements 1026 associated with different possible styles that can be selected to generate (or regenerate) the document template. Figure 10B The selectable example styles include "Expert Style Tone," "Casual Language," "Use of Metaphor," "Explicit Viewpoint," and "MBT Type: INTJ." The second user interface element 1027 can be configured to allow users to add new styles, enabling the document template to be regenerated based on the added styles.

[0201] For example, the second user interface screen 1020 includes a fourth portion 1028 associated with the document format of the generated document template. For example, the fourth portion 1028 of the second user interface screen 1020 includes multiple second user interface elements 1029 associated with different parts of the document structure, which can be selected and modified to generate (or regenerate) the document template. Figure 10B In the accompanying drawings, selectable example document sections include a "Title" section and a "Description" section. Other document sections may include a "Background" section and a "Conclusion" section. Each of the plurality of second user interface elements 1029 can be configured to be manipulated by the user, allowing different sections to be rearranged based on user input, thereby enabling the document template to be regenerated based on the rearranged document structure (format).

[0202] although Figure 10B Not shown, but the second user interface screen 1020 may also include an additional portion associated with the intent of the generated document template. For example, this additional portion of the second user interface screen 1020 may include multiple user interface elements associated with different possible intents that can be selected to generate (or regenerate) the document template. Examples of selectable intents include a "persuasive" intent, an "informative" intent, an "entertaining" intent, an "inspirational" intent, etc.

[0203] Figure 10C A visual depiction is shown illustrating how a document extractor application (which may be incorporated as part of a notebook application) according to one or more example embodiments of this disclosure is able to learn the format, style, and / or intent of a training document. Figure 10CAs shown, training document 1032 (“PRD for the product”) includes multiple sections 1034 for a PRD document, including an introduction, key user journey (CUJ) section, target audience section, problem statement section, and proposal section. Document extractor application 9100 can be configured to learn the document type of training document 10323 via selection of user interface elements 1036. Figure 10C As shown, the learned document portion 1038 for the PRD document includes the same portion as the training document 1032. The document extractor application 9100 can be configured to generate a PRD in response to a user-uploaded source document, wherein the generated PRD may have learned document type information obtained from the training document 1032. Figure 10C The format shown or a similar format.

[0204] For example, Figure 10D A fourth user interface screen (e.g., an output document generation user interface screen) of a document extractor application (which may be incorporated as part of a notebook application) according to one or more example embodiments of the present disclosure is shown.

[0205] exist Figure 10D In the fourth user interface screen 1040, a user interface (e.g., an output document generation user interface screen) is depicted that enables a user to create (generate) an output document based on one or more source documents that can be selected by the user, and according to a document template that can also be selected or provided to the user, for example, based on the content of these source documents. For example, in Figure 10D In the process, the user has selected multiple source documents 1042 (these source documents may correspond to notes that have been added to the draft area or notes section in the notebook application 3100).

[0206] In some implementations, the user can identify or select a specific document template and provide input that enables the document extractor application 9100 to generate an output document based on the selected source document and the selected document template. In some implementations, the document extractor application 9100 can be configured to analyze the content of the selected source documents and determine or suggest one or more document templates that may be suitable or applicable to these source documents.

[0207] For example, in Figure 10D In the fourth user interface screen 1040, a first user interface element 1044 is included, which is configured to create or generate a PRD based on a previously generated document template associated with the PRD document when selected.

[0208] For example, Figure 10EA fifth user interface screen (e.g., an output document user interface screen) of a document extractor application (which may be incorporated as part of a notebook application) according to one or more example embodiments of the present disclosure is shown.

[0209] exist Figure 10E In this context, the fifth user interface screen 1050 depicts a user interface (e.g., an output document user interface screen) including an output document generated by the document extractor application 9100 based on a selected source document and that document template (e.g., a PRD document template). For example, this output document could correspond to notes 1052 of the notebook application 3100. Figure 10E As shown, the output document can have a document structure or format consistent with the PRD document and can include similar sections, such as the introduction section 1054 and the CUJ section 1056. The document extractor application 9100 can be configured to generate content for each section based on the content of the source document, via one or more machine learning models, as described in the examples provided herein. For example, the output document can be stored on the computing device, transferred to another computing device, saved as a specific document file type in another document application (e.g., via a first user interface element 1058), or shared with another user (e.g., via a second user interface element 1059), etc.

[0210] Figure 11A A block diagram of an example computing system for organizing, managing, and creating content by implementing one or more machine learning-based models for source content and / or training content selected by a user, according to one or more example embodiments of the present disclosure, is depicted. System 1100 includes a user computing device 1102, a server computing system 1130, and a training computing system 1150 communicatively coupled via a network 1180.

[0211] Figure 11B A block diagram of an example computing apparatus for organizing, managing, and creating content by implementing one or more machine learning-based models for source content and / or training content selected by a user, according to one or more example embodiments of the present disclosure, is depicted.

[0212] Figure 11C A block diagram of an example computing apparatus for organizing, managing, and creating content by implementing one or more machine learning-based models for source content and / or training content selected by a user, according to one or more example embodiments of the present disclosure, is depicted.

[0213] User computing device 1102 (which may correspond to computing device 100) may be any type of computing device, such as, for example, a personal computing device (e.g., a laptop or desktop), a mobile computing device (e.g., a smartphone or tablet), a game console or controller, a wearable computing device, an embedded computing device, or any other type of computing device.

[0214] User computing device 1102 includes one or more processors 1112 and memory 1114. The one or more processors 1112 can be any suitable processing device (e.g., processor core, microprocessor, ASIC, FPGA, controller, microcontroller, etc.) and can be a single processor or multiple processors operatively connected. Memory 1114 can include one or more non-transitory computer-readable storage media, such as RAM, ROM, EEPROM, EPROM, flash memory devices, disks, etc., and combinations thereof. Memory 1114 can store data 1116 and instructions 1118, which are executed by processor 1112 to cause user computing device 1102 to perform operations.

[0215] In some implementations, the user computing device 1102 may store or include one or more machine learning models 1120 (e.g., large language models, sequence processing models, generative machine learning models, etc.). For example, the one or more machine learning models 1120 may be or otherwise include various machine learning models, such as neural networks (e.g., deep neural networks) or other types of machine learning models, including nonlinear and / or linear models. Example neural networks may include feedforward neural networks, recurrent neural networks (RNNs) (including recurrent neural networks based on long short-term memory (LSTM)), convolutional neural networks (CNNs), diffusion models, generative adversarial networks, or other forms of neural networks. Example neural networks may be deep neural networks. Some example machine learning models may utilize attention mechanisms, such as self-attention. For example, some example machine learning models may include multi-head self-attention models (e.g., Transformer models). Example machine learning models are referenced herein. Figures 1A to 10E It has been described.

[0216] In some implementations, the one or more machine learning models 1120 may be received from a server computing system 1130 via a network 1180, stored in memory 1114, and then used or otherwise implemented by the one or more processors 1112. In some implementations, the user computing device 1102 may implement multiple parallel instances of a single machine learning model (e.g., to perform parallel tasks across multiple instances of the machine learning model). In some implementations, the task is a generative task, and one or more machine learning models may be implemented to output content (e.g., document templates, output documents, etc.) in view of various inputs (e.g., queries, training documents, source documents, conditional parameters, etc.). More specifically, the machine learning models disclosed herein (e.g., including large language models, sequence processing models, generative machine learning models, etc.) may be implemented to perform various tasks related to input queries.

[0217] According to examples of this disclosure, a computing system may implement one or more sequence processing models 3120, 9120 as described herein to respond to or based on the query output for values ​​of the one or more conditions. The one or more sequence processing models 3120, 9120 may include one or more machine learning-enabled models configured to process and analyze sequence data and to process data appearing in a specific order or sequence, including time series data, natural language text, or any other data having a temporal or sequential structure.

[0218] According to examples in this disclosure, a computing system can implement one or more large language models 3130, 9130 to determine multiple variables based on the query. For example, the large language model may include a Bidirectional Encoder Representation (BERT) large language model from a Transformer. For example, the large language model can be trained to understand and process natural language. The large language model can be configured to extract information from the input (e.g., a query, training document, source document, etc.) to identify keywords, intent, and context within the input, thereby determining multiple variables for generating content. These variables may include latent variables representing the underlying structure of the language.

[0219] According to examples of this disclosure, a computing system may implement one or more generative machine learning models 3140, 9140 to generate various content (e.g., for generating outlines, summaries, responses to queries, document templates, output documents generated based on the document templates, etc.) having values ​​for one or more conditions. The one or more generative machine learning models 3140, 9140 may include deep neural networks or generative adversarial networks (GANs) to generate content having one or more features having values ​​for one or more conditions associated with those features. For example, the one or more generative machine learning models 3140, 9140 may include variational autoencoders, stable diffusion machine learning models, visual Transformers, neural radiation fields (NeRF), etc., to generate the content.

[0220] Alternatively, one or more machine learning models 1140 may be included in or otherwise stored and implemented by server computing system 1130, which communicates with user computing device 1102 according to a client-server relationship. For example, the one or more machine learning models 1140 may be implemented by server computing system 1130 as part of a web service (e.g., navigation service, word processing service, educational service, etc.). Therefore, one or more machine learning models 1120 may be stored and implemented at user computing device 1102, and / or one or more machine learning models 1140 may be stored and implemented at server computing system 1130.

[0221] User computing device 1102 may also include one or more user input components 1122 for receiving user input. For example, user input component 1122 may be a touch-sensitive component (e.g., a touch-sensitive display screen or touchpad) that is sensitive to the touch of a user input object (e.g., a finger or stylus). The touch-sensitive component may be used to implement a virtual keyboard. Other example user input components include a microphone, a conventional keyboard, or other means and methods through which a user may provide user input.

[0222] Server computing system 1130 (which may correspond to server computing system 300) includes one or more processors 1132 and memory 1134. The one or more processors 1132 may be any suitable processing device (e.g., processor core, microprocessor, ASIC, FPGA, controller, microcontroller, etc.) and may be a single processor or multiple processors operatively connected. Memory 1134 may include one or more non-transitory computer-readable storage media, such as RAM, ROM, EEPROM, EPROM, flash memory devices, disks, etc., and combinations thereof. Memory 1134 may store data 1136 and instructions 1138, which are executed by processor 1132 to cause server computing system 1130 to perform operations.

[0223] In some implementations, the server computing system 1130 includes one or more server computing devices or is otherwise implemented by one or more server computing devices. When the server computing system 1130 includes multiple server computing devices, such server computing devices may operate according to a sequential computing architecture, a parallel computing architecture, or some combination thereof.

[0224] As described above, the server computing system 1130 may store or otherwise include one or more machine learning models 1140. For example, the one or more machine learning models 1140 may be or otherwise include various machine learning models. Example machine learning models include neural networks or other multi-layer nonlinear models. Example neural networks may include feedforward neural networks, recurrent neural networks (RNNs) (including recurrent neural networks based on long short-term memory (LSTM), convolutional neural networks (CNNs), diffusion models, generative adversarial networks, or other forms of neural networks. Example neural networks may be deep neural networks. Some example machine learning models may utilize attention mechanisms, such as self-attention. For example, some example machine learning models may include multi-head self-attention models (e.g., Transformer models). Example machine learning models are referenced herein. Figures 1A to 10E It has been described.

[0225] User computing device 1102 and / or server computing system 1130 can train one or more machine learning models 1120 and / or 1140 via interaction with training computing system 1150, which is communicatively coupled via network 1180. Training computing system 1150 may be separate from server computing system 1130, or may be part of server computing system 1130.

[0226] The training computing system 1150 includes one or more processors 1152 and memory 1154. The one or more processors 1152 can be any suitable processing device (e.g., processor core, microprocessor, ASIC, FPGA, controller, microcontroller, etc.) and can be a single processor or multiple processors operatively connected. The memory 1154 can include one or more non-transitory computer-readable storage media, such as RAM, ROM, EEPROM, EPROM, flash memory devices, disks, etc., and combinations thereof. The memory 1154 can store data 1156 and instructions 1158, which are executed by the processor 1152 to cause the training computing system 1150 to perform operations. In some implementations, the training computing system 1150 includes one or more server computing devices or is otherwise implemented by one or more server computing devices.

[0227] The training computing system 1150 may include a model trainer 1160 that uses various training or learning techniques, such as, for example, error backpropagation, to train one or more machine learning models 1120 and / or 1140 stored at the user computing device 1102 and / or the server computing system 1130. For example, a loss function can be used to update one or more parameters of the model (e.g., based on the gradient of the loss function) through model backpropagation. Various loss functions can be used, such as mean squared error, likelihood loss, cross-entropy loss, hinge loss, and / or various other loss functions. Gradient descent techniques can be used to iteratively update the parameters across multiple training iterations.

[0228] In some implementations, performing error backpropagation may include performing truncated backpropagation through time. The model trainer 1160 may perform various generalization techniques (e.g., weight decay, dropout, etc.) to improve the generalization ability of the model being trained.

[0229] Specifically, model trainer 1160 can train the one or more machine learning models 1120 and / or 1140 based on a set of training data 1162. Training data 1162 can include, for example, various datasets that may be stored remotely or at the training computing system 1150. For example, in some implementations, example datasets used for training include large language training data corpora that provide one or more large language models with the ability to perform various language tasks. For example, the one or more large language models can be trained to perform tasks such as summarization, dialogue, simplification, and opposing viewpoints. In particular, the one or more large language models can be trained to process various outputs to generate language outputs. However, other datasets (e.g., datasets of images) (e.g., images obtained from external websites) can be utilized. In some implementations, the dataset can be limited to a specific genre or topic, a specific type of content (including image content, video, and text), a specific style or type of content (e.g., outlines, reports, presentations, spreadsheets, resumes, PRDs, etc.), and so on. In some implementations, the dataset can contain a variety of topics.

[0230] In some implementations, if the user has provided consent, training examples can be provided by the user computing device 1102. Therefore, in such implementations, the one or more machine learning-based models 1120 provided to the user computing device 1102 can be trained by the training computing system 1150 on user-specific data received from the user computing device 1102. In some cases, this process may be referred to as personalizing the model.

[0231] Model trainer 1160 includes computer logic for providing desired functionality. Model trainer 1160 can be implemented in hardware, firmware, and / or software that controls a general-purpose processor. For example, in some implementations, model trainer 1160 includes a program file stored on a storage device, loaded into memory, and executed by one or more processors. In other implementations, model trainer 1160 includes one or more sets of computer-executable instructions stored in a tangible computer-readable storage medium such as RAM, a hard disk, or an optical or magnetic medium.

[0232] Network 1180 can be any type of communication network, such as a local area network (e.g., intranet), a wide area network (e.g., the Internet), or some combination thereof, and can include any number of wired or wireless links. Generally, communication conducted through network 1180 can be carried via any type of wired and / or wireless connection, using various communication protocols (e.g., TCP / IP, HTTP, SMTP, FTP), encoding or format (e.g., HTML, XML), and / or protection schemes (e.g., VPN, Secure HTTP, SSL).

[0233] The machine learning models described in this specification can be used in a variety of tasks, applications, and / or use cases.

[0234] In some implementations, the input to the machine learning model of this disclosure can be text or natural language data. The machine learning model can process the text or natural language data to generate output. For example, the machine learning model can process natural language data to generate language-encoded output. For another example, the machine learning model can process text or natural language data to generate latent text embedding output. For another example, the machine learning model can process text or natural language data to generate translation output. For another example, the machine learning model can process text or natural language data to generate classification output. For another example, the machine learning model can process text or natural language data to generate text segmentation output. For another example, the machine learning model can process text or natural language data to generate semantic intent output. For another example, the machine learning model can process text or natural language data to generate quality-enhanced text or natural language output (e.g., text or natural language data of higher quality than the input text or natural language). For yet another example, the machine learning model can process text or natural language data to generate predictive output.

[0235] In some implementations, the input to the machine learning model of this disclosure can be speech data. The machine learning model can process the speech data to generate output. As an example, the machine learning model can process speech data to generate speech recognition output. As another example, the machine learning model can process speech data to generate speech translation output. As another example, the machine learning model can process speech data to generate latent embedding output. As another example, the machine learning model can process speech data to generate encoded speech output (e.g., encoded and / or compressed representations of speech data, etc.). As another example, the machine learning model can process speech data to generate quality-enhanced speech output (e.g., speech data of higher quality than the input speech data, etc.). As another example, the machine learning model can process speech data to generate text representation output (e.g., text representations of the input speech data, etc.). As another example, the machine learning model can process speech data to generate predicted output.

[0236] In some implementations, the input to the machine learning model of this disclosure can be sensor data. The machine learning model can process the sensor data to generate output. As an example, the machine learning model can process sensor data to generate recognition output. As another example, the machine learning model can process sensor data to generate prediction output. As another example, the machine learning model can process sensor data to generate classification output. As another example, the machine learning model can process sensor data to generate segmentation output. As another example, the machine learning model can process sensor data to generate visualization output. As another example, the machine learning model can process sensor data to generate diagnostic output. As another example, the machine learning model can process sensor data to generate detection output.

[0237] Figure 11A Example computing systems that can be used to implement various aspects of this disclosure are shown. Other computing systems may also be used. For example, in some implementations, user computing device 1102 may include model trainer 1160 and training data 1162. In such implementations, the one or more machine learning-based models 1120 may be trained and used locally at user computing device 1102. In some such implementations, user computing device 1102 may implement model trainer 1160 to personalize the one or more machine learning-based models 1120 based on user-specific data.

[0238] Figure 11BA block diagram of an example computing device, according to one or more example embodiments of the present disclosure, is depicted for organizing, managing, and creating content by implementing one or more machine learning-based models for training content and / or source content selected by a user. The computing device 1200 may be a user computing device or a server computing device.

[0239] The computing device 1200 includes multiple applications (e.g., applications 1 to N). Each application contains its own machine learning library and machine learning-enabled models. For example, each application may include machine learning-enabled models. Example applications include notebook applications as described herein, document extractor applications as described herein, text messaging applications, email applications, dictation applications, virtual keyboard applications, browser applications, social media applications, map applications, navigation applications, etc.

[0240] like Figure 11B As shown, each application can communicate with multiple other components of the computing device, such as, for example, one or more sensors, a context manager, a device state component, and / or additional components. In some implementations, each application can communicate with each device component using an API (e.g., a public API). In some implementations, the API used by each application is application-specific.

[0241] Figure 11C A block diagram of an example computing device according to one or more exemplary embodiments of the present disclosure is depicted. This example computing device is used to organize, manage, and create content by implementing one or more machine learning-based models for training content and / or source content selected by a user. The computing device 1300 may be a user computing device or a server computing device.

[0242] The computing device 1300 includes multiple applications (e.g., applications 1 to N). Each application communicates with a central intelligence layer. Example applications include a notebook application as described herein, a document extractor application as described herein, a text messaging application, an email application, a dictation application, a virtual keyboard application, a browser application, a map application, a social media application, a navigation application, and so on. In some implementations, each application may communicate with the central intelligence layer (and the models stored therein) using an API (e.g., a common API across all applications).

[0243] The central intelligence layer comprises multiple machine learning-engineered models. For example, such as... Figure 11CAs shown, a corresponding machine learning model can be provided for each application, and the corresponding machine learning model can be managed by a central intelligence layer. In other implementations, two or more applications can share a single machine learning model. For example, in some implementations, the central intelligence layer can provide a single model for all applications. In some implementations, the central intelligence layer is included within or otherwise implemented by the operating system of the computing device 1300.

[0244] The central intelligence layer can communicate with the central device data layer. The central device data layer can be a centralized data storage repository for the computing device 1300. For example... Figure 11C As shown, the central device data layer can communicate with multiple other components of the computing device, such as, for example, one or more sensors, a context manager, a device state component, and / or additional components. In some implementations, the central device data layer can communicate with each device component using an API (e.g., a private API).

[0245] As used herein, the terms "module" and "unit" are general terms that can refer to, but are not limited to, software or hardware components or devices that perform certain tasks, such as field-programmable gate arrays (FPGAs) or application-specific integrated circuits (ASICs). A module or unit may be configured to reside on an addressable memory medium and to execute on one or more processors. Thus, by way of example, a module or unit may include components (such as software components, object-oriented software components, class components, and task components), processes, functions, attributes, procedures, subroutines, program code segments, drivers, firmware, microcode, circuit systems, data, databases, data structures, tables, arrays, and variables. The functionality provided in components and modules / units may be combined into fewer components and modules / units or further divided into additional components and modules.

[0246] The aspects of the above-described example embodiments can be recorded in a non-transitory computer-readable medium, which includes program instructions for implementing various operations embodied by a computer. The medium may also include data files, data structures, etc., alone or in combination with the program instructions. Examples of non-transitory computer-readable media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical media such as CD-ROMs, Blu-ray discs, and DVDs; magneto-optical media such as optical discs; and other hardware devices specifically configured to store and execute program instructions, such as semiconductor memories, read-only memories (ROMs), random access memories (RAMs), flash memory, USB memories, etc. Examples of program instructions include machine code (such as that generated by a compiler) and files containing higher-level code that can be executed by a computer using an interpreter. The program instructions can be executed by one or more processors. The described hardware devices can be configured to act as one or more software modules to perform the operations of the above embodiments, and vice versa. Furthermore, non-transitory computer-readable storage media can be distributed among computer systems connected via a network, and computer-readable code or program instructions can be stored and executed in a distributed manner. In addition, non-transitory computer-readable storage media may also be embodied in at least one application-specific integrated circuit (ASIC) or field-programmable gate array (FPGA).

[0247] Each box in a flowchart can represent a unit, module, segment, or section of code, which includes one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions mentioned in the boxes may occur out of order. For example, two boxes shown consecutively may actually execute substantially concurrently (simultaneously), or these boxes may sometimes execute in reverse order, depending on the functions involved.

[0248] While this disclosure has been described with reference to various exemplary embodiments, each example is provided by way of explanation and not as a limitation thereof. Modifications, alterations, and equivalents of such embodiments will be readily apparent to those skilled in the art upon understanding the foregoing. Therefore, this disclosure does not exclude such modifications, alterations, and / or additions to the disclosed subject matter that will be readily understood by those of ordinary skill in the art. For example, features shown or described as part of one embodiment may be used with another embodiment to produce yet further embodiments. Therefore, this disclosure is intended to cover such modifications, alterations, and equivalents.

Claims

1. A computing device for generating output documents, comprising: One or more memories, the one or more memories being configured to store instructions; as well as One or more processors, the one or more processors being configured to execute the instructions to perform operations, the operations including: Receive multiple source documents, Receive input associated with a request to generate the output document based on the plurality of source documents and a specific document template, and The output document, which has the specific document template, is generated using one or more machine learning models.

2. The computing device as claimed in claim 1, wherein, The operation further includes: A graphical user interface is provided for presentation on a display device, the graphical user interface being configured to receive selections of the plurality of source documents.

3. The computing device as claimed in claim 1, wherein, The operation further includes: The intent associated with the output document is determined by one or more machine learning models based on the content associated with each of the plurality of source documents.

4. The computing device as claimed in claim 1, wherein, The operation further includes: The style associated with the output document is determined by one or more machine learning models based on the content of each of the plurality of source documents.

5. The computing device as claimed in claim 1, wherein, The operation further includes: The format associated with the output document is determined by one or more machine learning models based on the content of each of the plurality of source documents.

6. The computing device as claimed in claim 5, wherein, Determining the format associated with the output document by one or more machine learning models based on the content of each of the plurality of source documents includes identifying a plurality of headings for the corresponding sections of the output document.

7. The computing device as claimed in claim 1, wherein, The operation further includes: A graphical user interface is provided for presentation on a display device, the graphical user interface comprising a plurality of user interface elements, wherein each of the plurality of user interface elements corresponds to a different style that can be applied by the one or more machine learning models to generate the output document having the specific document template; and Receive selection of one or more user interface elements from the plurality of user interface elements. Generating the output document with the specific document template via the one or more machine learning models includes applying one or more styles corresponding to the selection of the one or more user interface elements among the plurality of user interface elements.

8. The computing device as claimed in claim 1, wherein, The operation further includes: A graphical user interface is provided for presentation on a display device, the graphical user interface comprising a plurality of user interface elements, wherein each of the plurality of user interface elements corresponds to a different heading that can be applied by the one or more machine learning models to generate portions of the output document having the specific document template; and Receive selection of one or more user interface elements from the plurality of user interface elements. Generating the output document with the specific document template via the one or more machine learning models includes applying one or more headings corresponding to the selection of the one or more user interface elements among the plurality of user interface elements to various parts of the output document.

9. The computing device of claim 1, further comprising: One or more databases are configured to store multiple generative machine learning models, each associated with a different document type. The operation further includes retrieving from the plurality of generative machine learning models at least one generative machine learning model associated with a document type indicated by the common content of the plurality of source documents.

10. The computing device of claim 1, wherein, The operation further includes: Receive a plurality of training documents associated with a document type, wherein each of the plurality of training documents includes one or more common features associated with the specific document template; and Train the one or more machine learning models to learn at least one of the intent, style, or format of the particular document template for the document type.

11. A computer-implemented method, comprising: Multiple source documents are received by a computing system comprising one or more processors and one or more machine learning-enabled models; The computing system receives input associated with a request to generate an output document based on the plurality of source documents and a specific document template; as well as The output document, having the specific document template, is generated via one or more machine learning models.

12. The computer-implemented method of claim 11, further comprising: The one or more machine learning models determine at least one of the style, intent, and format associated with the output document based on the content of each of the plurality of source documents.

13. The computer-implemented method of claim 11, further comprising: A graphical user interface is provided for presentation on a display device of the computing system, the graphical user interface comprising a plurality of user interface elements, wherein each of the plurality of user interface elements corresponds to a style, intent, or format that can be applied by the one or more machine learning models to generate the output document having the specific document template; and Receive selection of one or more user interface elements from the plurality of user interface elements. Generating the output document with the specific document template via the one or more machine learning models includes applying the style, intent, or format corresponding to the selection of the one or more user interface elements among the plurality of user interface elements.

14. The computer-implemented method of claim 11, further comprising: Store multiple generative machine learning models, each associated with a different document type, in one or more databases; as well as Retrieve from the plurality of generative machine learning models at least one generative machine learning model associated with a document type indicated by the common content of the plurality of source documents.

15. The computer-implemented method of claim 11, further comprising: Receive a plurality of training documents associated with a document type, wherein each of the plurality of training documents includes one or more common features associated with the specific document template; and Train the one or more machine learning models to learn at least one of the intent, style, or format of the particular document template for the document type.

16. A computing device for generating document templates, comprising: One or more memories, the one or more memories being configured to store instructions; as well as One or more processors, the one or more processors being configured to execute the instructions to perform operations, the operations including: Receive selections from multiple sample documents; Receive input for generating the document template based on the selection of the plurality of sample documents; and In response to receiving the input, one or more machine learning models are implemented to generate the document template based on at least one of the styles, formats, or intentions associated with the plurality of sample documents.

17. The computing device of claim 16, wherein, The operation further includes: Receive multiple source documents, Receive further input associated with the request to generate an output document based on the plurality of source documents and the document template, and The output document is generated based on the document template using one or more machine learning models.

18. The computing device of claim 17, wherein, The operation further includes: The one or more machine learning models determine at least one of the style, the format, and the intent associated with the plurality of sample documents based on the content and structure of each of the plurality of sample documents.

19. The computing device of claim 18, wherein, The operation further includes: A graphical user interface is provided for presentation on a display device, the graphical user interface comprising a plurality of user interface elements, wherein at least one of the plurality of user interface elements is selectable to change at least one of the style, the format, and the intent associated with the plurality of sample documents as determined by the one or more machine learning models; and Receive selection of one or more user interface elements from the plurality of user interface elements to change at least one of the style, format, and intent associated with the plurality of sample documents as determined by the one or more machine learning models. Generating the output document based on the document template via one or more machine learning models includes applying at least one of the style, the format, and the intent, which are altered based on the selection of one or more user interface elements among the plurality of user interface elements.

20. The computing device of claim 16, further comprising: One or more databases are configured to store multiple generative machine learning models, each associated with a different document type. The operation further includes retrieving from the plurality of generative machine learning models at least one generative machine learning model associated with a document type indicated by the common content of the plurality of sample documents.

Images