System and method for generating dynamically updated metadata using a real-time artificial intelligence model

Abstract

A system and method for generating dynamically updated metadata using a real-time artificial intelligence model. For example, the system can receive first metadata tag requirements for first metadata of a first media asset. Based on the first metadata tag requirements, the system can determine first metadata fields for the first metadata. The system can determine a first content population function for the first metadata fields. The system can generate first metadata using the first content population function. The system can generate a first media asset having the first metadata.

Description

Technical Field

[0001] Cross - reference to Related Applications

[0001] This application claims the benefit of priority of U.S. Patent Application No. 18 / 494,291, filed Oct. 25, 2023, which is a continuation of U.S. Patent Application No. 18 / 203,240, filed May 30, 2023. The content of the foregoing application is hereby incorporated by reference in its entirety.

Background Art

[0002] Background

[0002] Digital media refers to content or information that is created, distributed, and consumed using digital technologies. This includes various forms of media such as text, images, audio, video, and interactive elements that are generated and accessed via digital devices and platforms. Digital media has revolutionized the way content is created, shared, and / or consumed in many ways. Digital media has had a significant impact on industries such as entertainment, news, journalism, advertising, marketing, education, and communication. With the advent of the Internet and the widespread use of digital devices such as smartphones, tablets, and computers, digital media has become an indispensable part of daily life.

[0003]

[0003] While digital media offers many advantages, it also creates new technological challenges and problems. For example, the ease of creating and disseminating content on digital platforms has led to the proliferation of misinformation, rumors, and fake news. It can be difficult for users to distinguish between reliable information and false or misleading content. Furthermore, digital media often requires users to provide personal information, raising concerns about privacy and data security. There has been an increase in incidents of data breaches, unauthorized access, and misuse of personal information. In addition, digital media has made it easier to copy, distribute, and reproduce copyrighted content without proper permission. This has created significant challenges for content creators in protecting their intellectual property rights and revenue streams. [Overview of the project] [Means for solving the problem]

[0004] overview

[0004] In consideration of the technical challenges and problems specific to digital media described herein, the systems and methods described herein offer improvements to how digital media are produced, shared, and / or consumed. In particular, the systems and methods described herein enumerate novel metadata architectures and uses. For example, the systems and methods cite the use of generating dynamically updated metadata using a real-time artificial intelligence model. Dynamically updated metadata can mitigate the novel technical challenges and problems described above by automatically updating itself to reflect real-time changes in the content it describes and its relationships with other content.

[0005]

[0005] For example, in conventional systems, the metadata of a media asset, whether public or private, is static unless it is forcibly updated or modified by a user. As a result, the metadata lacks robustness, which limits its usefulness in learning about the quality, origin, use, rights, etc., associated with that media asset. One solution to overcome this lack of robustness is to "embed" these characteristics into the digital asset using blockchain and other ledger technologies, but the use of such technologies has given rise to many new problems and challenges.

[0006]

[0006] In contrast to this approach, the system and method can dynamically update metadata in real time by continuously mining public data from one or more databases using artificial intelligence models, including but not limited to machine learning and deep learning, specifically large-scale language models (LLMs). The updated metadata can reflect changes in the underlying media assets and their relationships with other media assets. For example, the metadata can provide a probability / confidence of how accurate the metadata is, whether (and / or to what extent) the media asset is similar to other media assets, and / or how (and / or by whom) the media asset was used. Furthermore, this solution overcomes the challenges with respect to metadata mentioned above and does not introduce technical problems inherent in blockchain technology.

[0007]

[0007] To achieve these technical advantages, the system and method describe the use of metadata that includes a content population function within the metadata itself. The content population function, which may be driven by an artificial intelligence model, can monitor input (e.g., data parsed from one or more databases) and update the content of the metadata (e.g., a description of the media asset, usage information about the media asset, etc.). In doing so, the system can maintain metadata relevant to the current situation, and / or metadata describing relationships and / or data that may not have existed at the time the metadata and / or the media asset were created.

[0008]

[0008] In some embodiments, systems and methods for generating dynamically updated metadata using a real-time artificial intelligence model are described. For example, the system can receive first metadata tag requirements relating to first metadata of a first media asset. Based on the first metadata tag requirements, the system can determine first metadata fields relating to the first metadata. The system can determine a first content population function relating to the first metadata fields. The system can generate first metadata using the first content population function. The system can generate a first media asset having the first metadata.

[0009]

[0009] Various other aspects, features and advantages of the present invention will become apparent through the detailed description of the invention and the drawings supplementary to this specification. It should be understood that both the above summary and the following detailed description are examples and do not limit the scope of the invention. As used herein and in the claims, the singular forms “a,” “an,” and “the” include a plurality of referenced objects unless the content clearly indicates otherwise. In addition, as used herein and in the claims, the term “or” means “and / or” unless the content clearly indicates otherwise. [Brief explanation of the drawing]

[0010] Brief explanation of the drawing [Figure 1]

[0010] An example is provided that is useful in describing a user interface for presenting multiple media assets, according to one or more embodiments. [Figure 2]

[0011] Examples are provided to help illustrate pseudocode for dynamically updated metadata, according to one or more embodiments. [Figure 3]

[0012] This document presents a machine learning model architecture for facilitating dynamically updated metadata, in one or more embodiments. [Figure 4]

[0013] One or more embodiments of a system for facilitating dynamically updated metadata are shown. [Figure 5]

[0014] A flowchart of the steps involved in dynamically updated metadata, according to one or more embodiments, is shown. [Figure 6]

[0015] A flowchart is shown illustrating the use of an artificial intelligence model to generate dynamically updated metadata, according to one or more embodiments. [Modes for carrying out the invention]

[0011] Detailed description of the drawing

[0016] In the following description, many specific details are given for illustrative purposes in order to provide a complete understanding of the embodiments of the present invention. However, it will be understood by those skilled in the art that embodiments of the present invention can be carried out without these specific details or with equivalent configurations. In other cases, well-known structures and devices are shown in block diagram form to avoid unnecessarily obscuring embodiments of the present invention.

[0012]

[0017] While the present invention has been described in detail for illustrative purposes based on what is currently considered to be the most practical and preferred embodiment, it should be understood that such details are for that purpose only, and the present invention is not limited to the disclosed embodiments, but rather intended to encompass additional claims, modifications, and equivalent configurations. For example, it should be understood that the present invention intends, wherever possible, to combine one or more features of any embodiment with one or more features of any other embodiment.

[0013]

[0018] Figure 1 shows an example that helps describe a user interface presenting multiple media assets in one or more embodiments. A system and method for generating dynamically updated metadata using a real-time artificial intelligence model is described. For example, Figure 1 shows user interface 100. Where used herein, “user interface” may include human-computer interaction and communication on a device and may include the appearance of a display screen, keyboard, mouse, and desktop. For example, a user interface may include a method for a user to interact with an application or website.

[0014]

[0019] User interface 100 includes media assets 102 and 104. Where used herein, “media asset” should be understood to mean electronically consumable content such as internet content (e.g., streaming content, downloadable content, webcasts, etc.), video clips, audio, content information, pictures, rotated images, documents, playlists, websites, articles, books, ebooks, blogs, advertisements, chat sessions, social media content, applications, games, and / or any other media or multimedia, and / or combinations thereof. Media assets may include digital media (or digital assets) which are digitized content that can be transmitted over the internet or computer networks. Media assets may also include augmented reality (AR), virtual reality (VR), mixed reality (MR), and / or other extended reality (XR). Media assets may be recorded, played back, displayed or accessed by a user device, but may also be part of a live performance or other user-generated content. Furthermore, user-generated content may include content created and / or consumed by the user. For example, user-generated content may include content created by another user but consumed and / or published by that user. It should be noted that embodiments describing one type of media (e.g., video) may also apply to other types of media (e.g., audio).

[0015]

[0020] Media assets 102 and 104 may each contain metadata. For example, media asset 102 may contain first metadata, and media asset 104 may contain second metadata. For example, metadata refers to data that provides information about other data. In some embodiments, the system may receive first metadata tag requirements regarding the metadata of a media asset. In some embodiments, the metadata tag requirements may indicate the purpose of the metadata tag. The purpose of the metadata tag may be to provide context, description, and / or structure to the main data it represents. For example, the metadata tag requirements may describe and identify various aspects of the data, such as title, author, creation date, file format, and / or size. That is, the metadata tag requirements may provide important information about the content and characteristics of the data.

[0016]

[0021] Additionally or alternatively, metadata tag requirements can improve data discoverability by enabling efficient searching and indexing. Therefore, metadata tag requirements can include keywords, tags, and / or classifications, making it easier to find and retrieve specific data within larger collections. Additionally or alternatively, metadata tag requirements can establish context and / or relationships between different data. Thus, metadata tag requirements can indicate source, location, and / or versioning information, enabling users (and / or other systems) to understand the origin and relevance of data. Additionally or alternatively, metadata tag requirements may play a crucial role in long-term data preservation and management. Therefore, metadata tag requirements may include information about the data's origin, rights, access restrictions, and archiving procedures, ensuring that data remains accessible, authentic, and / or usable over time. Additionally or alternatively, metadata tag requirements can facilitate data integration by providing information about data format, structure, and schema. Thus, metadata tag requirements can enable the joining, analysis, and consistent interpretation of different datasets, and enable interoperability between different systems and / or applications. Additionally or alternatively, metadata tag requirements can support data quality initiatives by capturing information about data quality standards, validation rules, and data lineage. Therefore, metadata tag requirements can help organizations maintain data integrity, track changes, and / or enforce data governance policies. Additionally or alternatively, metadata tag requirements may include security-related information such as access controls, encryption details, and data sensitivity classifications. Therefore, metadata tag requirements can help ensure appropriate security measures are applied and protect sensitive information. Additionally or alternatively, metadata tag requirements can enhance data understanding, accessibility, and management throughout its entire lifecycle.Therefore, metadata tagging requirements enable effective data discovery, integration, storage, and governance, potentially benefiting a wide range of areas, including libraries, archives, databases, digital media, scientific research, and / or other fields.

[0017]

[0022] The system can determine a first metadata field relating to the first metadata based on the first metadata tag requirements. In some embodiments, the metadata field may include metadata components or attributes that provide descriptive information about specific data or content. Metadata fields may be used by the system to classify, organize, and manage data in various systems such as databases, content management systems, and digital libraries. Metadata fields may function as labels or tags that can be associated with data to provide additional context and information about that data. These fields may contain various types of information. For example, a metadata field may include a descriptive field that describes the content, including title, author, creation date, subject, keywords, and / or summary. Additionally or alternatively, a metadata field may include a structural field that defines the structure or organization of the data, such as chapters, sections, and subsections within a document. Additionally or alternatively, a metadata field may include a management field that captures management information such as file format, file size, permissions, access restrictions, and / or version number. Additionally or alternatively, metadata fields may include technical fields that record technical details about the data, such as resolution, file type, encoding, color space, and the software used to create or modify the content. Additionally or alternatively, metadata fields may include rights management fields that specify copyright information, license terms, and / or usage restrictions.

[0018]

[0023] In some embodiments, metadata fields can be predefined and standardized to ensure consistency and interoperability across different systems and applications. The fields can enable efficient search, retrieval, and management of data by allowing users to filter and sort based on specific criteria. Metadata fields play an important role in organizing and understanding large amounts of data, improving data discoverability, and facilitating effective data governance.

[0019]

[0024] In some embodiments, the system can determine the visual representation of data and apply consistent naming rules, default values, and semantics to one or more fields within the model. These naming rules, default values, and semantics of one or more fields within the model can then be used by the system to generate recommendations for applications. For example, each field can correspond to a category of criteria, characteristics, and / or options. The system can use the field identifier to identify the type of criteria being input. For example, the system can compare the field identifier to a field database (e.g., a lookup table database listing the content corresponding to the field and / or the characteristics of the content) to identify content for recommendations.

[0020]

[0025] Each field can correspond to criteria regarding specific information and / or information about specific characteristics of the content. Alternatively or additionally, each field can provide a given function. This function can be a function executed locally (e.g., a function executed on a local device), or this function can be a function executed remotely. In some embodiments, the function can include additional information and / or links to other applications that are accessible and / or available locally or remotely. In some embodiments, the field can be represented by text and / or graphical information.

[0021]

[0026] In some embodiments, the system can detect information regarding the fields of an application (e.g., metadata or other information describing the fields). For example, the information can describe the purpose, function, origin, creator, developer, system requirements (including required format and / or capabilities), author, recommended uses, and / or approved users, etc. The information may be expressed in a human-readable and / or computer-readable language and may not be perceivable by a user browsing the user interface 100. These fields can be used by the system to match against criteria and / or other information submitted by the user and / or content provider. For example, in some embodiments, the system can receive content and / or criteria from multiple users and / or providers. In some embodiments, these criteria can describe the content and / or describe processing actions related to a given piece of content. For example, a first resource provider can input criteria regarding the price of content (e.g., a given digital asset) and / or criteria regarding a first set of delivery conditions for the content. A second provider can input criteria regarding a second set of delivery conditions for the content. Then, the user can input criteria regarding acceptable delivery conditions for the content. The system can match each received criterion against a content field identifier (e.g., a value that uniquely identifies the content and / or characteristics related to the content). Then, the system can make recommendations related to the content. For example, the system can recommend content having a first set of delivery conditions to the user (since these are superior to the second set of delivery conditions).

[0022]

[0027] A field may include a field identifier and / or field characteristics associated with a particular type of data. For example, field characteristics could be information (e.g., ordering, heading information, title, description, rating information, source code data (e.g., HTML, source code header, etc.), genre or category information, subject information, author / actor information, logo data, or other identifiers of the content provider), media format, file type, object type, objects appearing in the content (e.g., product placement, advertisement, keywords, context), or any other appropriate information used to distinguish one section from another. In some embodiments, field characteristics may also be human-readable text. Field characteristics may be determined to indicate fields (or content related to values ​​entered in the fields) of interest to the user based on a comparison of the field characteristics with user profile data about the user. In some embodiments, a field may include the name of metadata.

[0023]

[0028] The system can determine a first content population function for a first metadata field. In some embodiments, the first metadata may be input to a content population function. For example, the content population function may be a computer code function, which is referred to herein simply as a content-generating function. Where used herein, a function can be a self-contained block of code that performs a particular task or a set of related tasks. A function may accept inputs known as arguments or parameters and be designed to produce an output or perform a particular action based on those inputs. Functions can be called or invoked from other parts of a program, enabling modular and structured code construction. Functions can improve the readability, reusability, and maintainability of the code. Depending on the programming language and the specific requirements of the function, a function may have different types of return values, or it may have no return values ​​at all. A function may also perform additional actions, such as modifying variables outside its scope or printing information to the console.

[0024]

[0029] In some embodiments, a function may form the basis of a dynamic meta tag, which, as referred herein, may be a type of meta tag that changes based on specific conditions or variables, such as the content of a web page or the search query used to find the page. Dynamic meta tags may be used by systems to provide search engines and users with more relevant specific information. For example, a dynamic meta title tag may include the name of a specific product or service offered on the web page, or the location of a business. Tags can help pages rank higher in search results for relevant queries and attract more targeted traffic.

[0025]

[0030] The system can generate first metadata using a first content population function. In some embodiments, the system can generate metadata using a content population function using manual and / or automated input. For example, by identifying a specific content population function to include, the system can generate metadata. The content itself may be created by a human author and / or generated using automated methods such as a natural language processing (NLP) model like GPT-4.0.

[0026]

[0031] In some embodiments, metadata can be automatically generated or suggested based on a content population function. This can be achieved using techniques such as NLP, machine learning, or keyword analysis. For example, an algorithm may analyze content and suggest that a content population function generate metadata code (using, for example, relevant tags or categories based on its text content, structure, or context). In some embodiments, the content population function may be based on artificial intelligence models (collectively referred to herein as artificial intelligence models, machine learning models, or simply models), including but not limited to machine learning and deep learning.

[0027]

[0032] The system can generate a first media asset having first metadata. In some embodiments, the system can generate a media asset with metadata by attaching the metadata to the media asset along with a content population function. Once the media asset is ready, the associated metadata is assigned to it. This can be done manually by a human editor or automated by a script or algorithm. For example, the editor can input the metadata into a content management system (CMS) or provide it as input to an automated process.

[0028]

[0033] A system can attach metadata in various ways depending on the specific context and purpose. For example, a system can embed metadata in media assets and / or the media asset files themselves. For instance, in digital photographs, metadata such as the camera manufacturer and model, date and time of capture, GPS coordinates, and other relevant details can be stored within the image file using a standardized format such as EXIF ​​(Exchangeable Image File Format). Similarly, audio and video files may contain metadata such as artist name, album title, and playback time using formats such as ID3 tags or MPEG-7. Additionally or alternatively, a system can generate file system metadata. For example, operating systems often provide a mechanism for attaching metadata to files in the file system. This metadata may include attributes such as file name, file size, creation and modification dates, permissions, and other file-specific details. The file system manages this metadata together with the files themselves. Additionally or alternatively, a system can generate database metadata. For example, in the context of a relational database, metadata is stored in special system tables or dictionaries. These metadata records describe the structure and characteristics of the database, including tables, columns, indexes, relationships, and constraints. The database management system (DBMS) uses this metadata to optimize query execution and ensure data integrity.

[0029]

[0034] Additionally or alternatively, systems can generate document metadata. Document formats such as Microsoft Word, PDF, or HTML often provide fields or properties into which metadata can be entered manually or automatically. This metadata may include information such as the document's title, author, subject, keywords, creation date, and revision history. Users can usually access this metadata from the document properties or properties panel of the respective software. Additionally or alternatively, systems can generate web page metadata. For example, websites use metadata to provide additional information about web pages. One widely used standard is meta tags within the HTML code of a web page. These tags may include metadata such as page title, description, keywords, author, and language. Web crawlers and search engines use this metadata to index and display search results. Additionally or alternatively, systems can generate network protocol metadata. For example, some network protocols, such as HTTP, can attach metadata to content during transmission. For example, HTTP headers can carry metadata about content type, encoding, cache directives, authentication, etc. These headers are sent along with the content itself as part of the communication between the client and the server.

[0030]

[0035] Figure 2 shows an example that helps illustrate pseudocode for dynamically updated metadata in one or more embodiments. For example, code 200 and / or code 250 may be used in some embodiments to generate dynamically updated metadata. That is, code 200 and / or code 250 may be used to generate media assets that have metadata (e.g., meta tags) that include a content population function. For example, a dynamic meta tag is a type of meta tag that changes based on specific conditions or variables, such as the content of a web page or a search query used to find the page. Dynamic meta tags may be used to provide search engines and users with more relevant and specific information. For example, a dynamic meta title tag may include the name of a specific product or service offered on the web page, or the location of a business, which may help the page rank higher in search results for relevant queries and attract more targeted traffic.

[0031]

[0036] As described herein, the system can generate dynamic meta tags that may be automatically generated by a CMS or other web development tool based on data from a web page or database. The system can also allow for the manual creation of meta tags using a server-side scripting language such as PHP, which can pull data from a database or other source and populate the meta tag fields. Creating dynamic meta tags requires dynamically updating the content of the meta tags based on specific conditions or events, using a programming language such as JavaScript or PHP. An example of dynamic meta tags based on PHP is reflected in Code 200. As shown in Code 200, the system sets variables for page title, description, and keywords based on the page content or database data. These variables are then output as meta tags in the head section of the page's HTML using an echo statement.

[0032]

[0037] Code 250 demonstrates how to dynamically update dynamic meta tags using JavaScript. For example, the HTML page in Code 250 contains default meta descriptions for a section of that page. Once the page has finished loading, the JavaScript code uses the `document.querySelector()` method to select a meta description tag and the `setAttribute()` method to update the tag's content attribute.

[0033]

[0038] By using Code 200, Code 250, and / or other codes, the system can automatically tag content for digital asset management using various techniques (based on artificial intelligence modes including, for example, machine learning, computer vision, deep learning, LLM, and NLP), a subfield of artificial intelligence that focuses on the interaction between computer and human language. The system may also use automated metadata extraction, which involves extracting relevant information from file properties such as file name, date, and file type, and / or integration with content management systems.

[0034]

[0039] To perform automated metadata extraction, the system can automatically extract metadata from digital files such as documents, images, videos, and audio files by using text analysis to analyze the file's content for metadata extraction, image analysis to analyze the visual content of images for metadata extraction, and / or audio and video analysis to analyze the file's audio and visual content for metadata extraction.

[0035]

[0040] In some embodiments, Code 200 and / or Code 250 may be included in a self-executing program or smart contract (e.g., on a blockchain). For example, a meta tag can perform a blockchain operation. Where used herein, “blockchain operation” may include any operation including and / or related to the blockchain and blockchain technology. For example, a blockchain operation may include performing a transaction, querying a distributed ledger, generating additional blocks for the blockchain, sending communication-related non-fungible tokens, performing encryption / decryption, exchanging public / private keys, and / or other operations related to the blockchain and blockchain technology. In some embodiments, a blockchain operation may include creating, modifying, discovering, and / or executing a smart contract or program stored on the blockchain. For example, a smart contract may include a program stored on the blockchain that is executed (e.g., automatically, without the involvement of an intermediary or time loss) when one or more predetermined conditions are met. In some embodiments, a blockchain operation may include creating, modifying, exchanging, and / or examining tokens (e.g., digital blockchain-specific assets) that include non-fungible tokens. Non-fungible tokens may include tokens associated with goods, services, smart contracts, and / or other content that are verified by blockchain technology and can be stored using blockchain technology.

[0036]

[0041] Figure 3 shows an artificial intelligence model architecture for facilitating dynamically updated metadata, according to one or more embodiments. For example, the system may include one or more artificial intelligence models, architectures, and / or data preparation steps for generating meta tags. The system can determine which artificial intelligence models to use for one or more decisions used herein to generate decisions and / or metadata (e.g., how to tag content, how to tag media assets, how to interpret user-selected criteria, how to tag metadata requirements, and / or how to interpret other criteria). The system can select the artificial intelligence model (e.g., from multiple artificial intelligence models) that best provides the most accurate results. For example, the system can select from various ensemble architectures featuring one or more models trained (e.g., in parallel) to provide the most accurate results.

[0037]

[0042] Model 300 represents an artificial neural network. Model 300 includes an input level 302. The input level 302 can receive data related to known metadata tag requirements, metadata fields, content population functions, etc. Model 300 also includes one or more hidden layers (e.g., hidden layers 304 and 306). Model 300 may be based on a large collection of neural units (or artificial neurons). Model 300 roughly mimics how a biological brain works (e.g., via a large cluster of biological neurons connected by axons). Each neural unit in Model 300 may be connected to many other neural units in Model 300. Such connections can enforce or suppress their influence on the activation state of the connected neural units. In some embodiments, each individual neural unit may have an additive function that combines all the values ​​of its input together. In some embodiments, each connection (or the neural unit itself) may have a threshold function that a signal must pass before propagating to other neural units. Model 300 may be self-learning and trained rather than explicitly programmed, and may perform significantly better in certain areas of problem-solving compared to conventional computer programs. During training, output layer 308 may correspond to classifications of Model 300 (e.g., contacts awaiting assignment to an agent), and inputs known to correspond to such classifications may be input to input level 302. In some embodiments, Model 300 may include multiple layers (e.g., signal paths passing from front layer to back layer). In some embodiments, backpropagation techniques may be utilized by Model 300, where forward stimuli are used to reset weights on the "front" neural units. In some embodiments, stimuli and inhibitors of Model 300 may be more free-flowing, and connections interact in a more chaotic and complex manner. Model 300 also includes output layer 308.During testing, the output layer 308 may indicate whether a given input corresponds to a classification in model 300 (e.g., known metadata tag requirements, metadata fields, content population function, etc.).

[0038]

[0043] Figure 3 also includes Model 350, which is a convolutional neural network. A convolutional neural network is an artificial neural network characterized by one or more convolutional layers. The convolutional layers extract features from the input. The convolution preserves the relationships between the input data by learning the features using partitions of the input data. As shown in Model 350, the input layer 352 may proceed to convolutional blocks 354, 356 before outputting to the convolutional output 360. In some embodiments, Model 350 may itself serve as an input to Model 300.

[0039]

[0044] In some embodiments, Model 350 can implement an inverted residual structure where the inputs and outputs of a residual block (e.g., block 354) are thin bottleneck layers. The residual layer can feed to the next layer and directly to one or more downstream layers. The bottleneck layer (e.g., block 358) is a layer with fewer neural units compared to the previous layer. Model 350 can use the bottleneck layer to obtain a reduced-dimensional representation of the input. An example of this is the use of an autoencoder with a bottleneck layer for nonlinear dimensionality reduction. In addition, Model 350 may remove nonlinearity in a narrow layer (e.g., block 358) to maintain expressiveness. In some embodiments, the design of Model 350 may also be guided by a computational complexity metric (e.g., the number of floating-point operations). In some embodiments, Model 350 may increase the feature map dimension in all units to include as many locations as possible, instead of abruptly increasing the feature map dimension in the neural units performing downsampling. In some embodiments, Model 350 may reduce the depth and increase the width of the residual layer in the downstream direction.

[0040]

[0045] Figure 4 shows a system for facilitating dynamically updated metadata in one or more embodiments. As shown in Figure 4, the system 400 may include a server 422 and a user terminal 424 (which may correspond to a personal computer in some embodiments). Note that although shown as a server and a personal computer in Figure 4, the server 422 and the user terminal 424 may be any computing device, including but not limited to laptop computers, tablet computers, handheld computers, and other computer equipment (e.g., servers) including “smart,” wireless, wearable, and / or mobile devices. Figure 4 also includes a cloud component 410. The cloud component 410 may alternatively be any computing device as described above, and may include any type of mobile terminal, fixed terminal, or other device. For example, the cloud component 410 may be implemented as a cloud computing system and may feature one or more component devices. Note that the system 400 is not limited to three devices. Users can interact with each other, with one or more servers, or with other components of System 400, for example, using one or more devices. While one or more operations are described herein as being performed by specific components of System 400, it should be noted that in some embodiments, these operations may be performed by other components of System 400. For example, while one or more operations are described herein as being performed by components of Server 422, these operations may in some embodiments be performed by components of Cloud Component 410. In some embodiments, the various computers and systems described herein may include one or more computing devices programmed to perform the described functions. Additionally or alternatively, multiple users may interact with System 400 and / or one or more components of System 400.For example, in one embodiment, a first user and a second user may interact with the system 400 using two different components.

[0041]

[0046] With respect to the components of Server 422, User Terminal 424, and Cloud Component 410, each of these devices can receive content and data via input / output (hereinafter "I / O") paths. Each of these devices may also include a processor and / or control circuit for sending and receiving commands, requests, and other appropriate data using the I / O paths. The control circuit may comprise any appropriate processing circuits, memory circuits, and / or input / output circuits. Each of these devices may also include a user input interface and / or user output interface (e.g., a display) for use when receiving and displaying data. For example, as shown in Figure 4, both Server 422 and User Terminal 424 include a display for displaying data (e.g., as shown in Figure 1).

[0042]

[0047] In addition, since the server 422 and user terminal 424 are shown as a touchscreen smartphone and a personal computer, their displays also function as user input interfaces. Note that in some embodiments, the devices may not have a user input interface or a display, and instead may receive and display content using other devices (e.g., a dedicated display device such as a computer screen and / or a dedicated input device such as a remote control, mouse, or voice input). In addition, devices within the system 400 may run applications (or other suitable programs). The applications may cause the processor and / or control circuits to perform actions related to recommending content. While some embodiments are described herein in particular with respect to artificial intelligence models, note that in other embodiments, analytical models based on other predictive statistics may be used instead of, or in addition to, artificial intelligence models.

[0043]

[0048] Each of these devices may also include memory in the form of electronic storage. Electronic storage may include non-temporary storage media that electronically store information. The electronic storage media of electronic storage may include (i) system storage provided integrally with a server or client device (e.g., substantially inremovable), or (ii) removable storage that is removablely connected to a server or client device via, for example, a port (e.g., a USB port, a FireWire port, etc.) or a drive (e.g., a disk drive, etc.). Electronic storage may include one or more optically readable storage media (e.g., optical discs, etc.), magnetically readable storage media (e.g., magnetic tape, magnetic hard drives, floppy drives, etc.), charge-based storage media (e.g., EEPROM, RAM, etc.), solid-state storage media (e.g., flash drives, etc.), and / or other electronically readable storage media. Electronic storage may include one or more virtual storage resources (e.g., cloud storage, virtual private networks, and / or other virtual storage resources). Electronic storage can store information determined by software algorithms, processors, information obtained from servers, information obtained from client devices, or other information that enables functions as described herein.

[0044]

[0049] Figure 4 also includes communication paths 428, 430, and 432. Communication paths 428, 430, and 432 may include the Internet, cellular networks, mobile voice or data networks (e.g., 5G or LTE networks), cable networks, public switched telephone networks, or other types of communication networks, or combinations of communication networks. Communication paths 428, 430, and 432 may include one or more communication paths separately or together, such as satellite paths, fiber optic paths, cable paths, paths supporting Internet communications (e.g., IPTV), free-space connections (e.g., for broadcast or other radio signals), or any other suitable wired or wireless paths, or combinations of such paths. A computing device may include additional communication paths linking multiple hardware, software, and / or firmware components working together. For example, a computing device may be implemented by a cloud of computing platforms working together as a computing device.

[0045]

[0050] The cloud component 410 may be a database (table or graph) configured to store user data of the system. For example, the database may contain data on media assets, metadata, and / or correlations between them that the system has collected through previous interactions, both active and passive. Alternatively or additionally, the system may function as a clearinghouse for multiple sources of information about data, available resources, and / or other content. For example, one or more of the cloud components 410 may include microservices and / or components thereof. In some embodiments, a microservice may be a collection of applications, each collecting one or more of a plurality of variables.

[0046]

[0051] The cloud component 410 may include a model 402, which may be an artificial intelligence model and / or another artificial intelligence model (as shown in Figure 3). Model 402 may take an input 404 and provide an output 406. The input may include multiple datasets, such as a training dataset and a test dataset. Each of the multiple datasets (e.g., input 404) may include a subset of data related to user data, original content, and / or alternative content. In some embodiments, the output 406 may be fed back to model 402 as input to the training model 402. For example, the system may receive a first labeled feature input, which is labeled with known descriptions (e.g., known metadata tag requirements, metadata fields, content population functions, etc.) for the first labeled feature input (e.g., a feature input based on labeled training data). The system may then train a first artificial intelligence model to classify the first labeled feature input with the known descriptions.

[0047]

[0052] In another embodiment, Model 402 can update its configuration (e.g., weights, biases, or other parameters) based on an evaluation of its prediction (e.g., output 406) and reference feedback information (e.g., user indication of accuracy, reference label, or other information). In another embodiment where Model 402 is a neural network, connection weights can be adjusted to match the difference between the neural network's prediction and the reference feedback. In further use cases, one or more neurons (or nodes) of the neural network may require their respective errors to be transmitted backward through the neural network (e.g., backpropagation of errors) to facilitate the update process. The update of connection weights may reflect, for example, the magnitude of the error propagated backward after forward propagation is complete. In this way, for example, Model 402 can be trained to produce better predictions.

[0048]

[0053] In some embodiments, Model 402 may include an artificial neural network. In such embodiments, Model 402 may include an input layer and one or more hidden layers. Each neural unit of Model 402 may be connected to many other neural units of Model 402. Such connections can enforce or suppress their influence on the activation state of the connected neural units. In some embodiments, each individual neural unit may have an additive function that combines all the values ​​of its input. In some embodiments, each connection (or the neural unit itself) may have a threshold function that a signal must exceed before propagating to other neural units. Model 402 may be self-learning and trained rather than explicitly programmed, and may perform significantly better in certain areas of problem-solving compared to conventional computer programs. During training, the output layer of Model 402 may correspond to a classification of Model 402, and inputs known to correspond to that classification may be input to the input layer of Model 402 during training. During testing, inputs without known classifications may be input to the input layer, and the determined classification may be output.

[0049]

[0054] In some embodiments, Model 402 may include multiple layers (e.g., a signal path traversing from a front layer to a back layer). In some embodiments, Model 402 may utilize a backpropagation technique, where a forward stimulus is used to reset weights on the "front" neural unit. In some embodiments, the stimuli and inhibitors of Model 402 may be more free-flowing, and the connections interact in a more chaotic and complex manner. During testing, the output layer of Model 402 may indicate whether a given input corresponds to a classification of Model 402 (e.g., an incident).

[0050]

[0055] For example, in some embodiments, the system may train an artificial intelligence model (e.g., an artificial neural network) to detect known descriptions based on feature inputs. For example, the system may receive user data (including, for example, the variables and categories of variables described herein). The system may then generate a set of feature inputs based on the training data. For example, the system may generate a first feature input based on training data that includes user data corresponding to a first known error (or error likelihood). The system may label the first feature input with a first known description (e.g., label the data as corresponding to a classification of descriptions).

[0051]

[0056] For example, in some embodiments, the system can train an artificial intelligence model (e.g., an artificial neural network) to determine known metadata tag requirements, metadata fields, content population functions, etc. For example, the system can receive criteria (e.g., known metadata tag requirements, metadata fields, content population functions, etc.). The system can then generate a set of feature inputs based on the criteria. For example, the system can generate feature inputs based on training data containing content corresponding to the model's interpretation of a user description, and the system can determine a response (e.g., known metadata tag requirements, metadata fields, content population functions, etc.).

[0052]

[0057] The system can then train an artificial intelligence model to detect a first known content based on a first labeled feature input. The system can also train an artificial intelligence model (e.g., the same or a different artificial intelligence model) to detect a second known content based on a second labeled feature input. For example, the training process may involve initializing several random values ​​for each of the training matrices (e.g., the artificial intelligence model) and attempting to predict the output of the input features using the initial random values. Initially, the model's error will be large, but by comparing the model's predictions to the correct output (e.g., a known classification), the model can adjust its weights and bias values ​​until it provides the desired prediction.

[0053]

[0058] In some embodiments, the system may use one or more modeling techniques, including supervised modeling. Such supervised machine learning techniques, such as linear or nonlinear regression, including neural networks and support vector machines, may be used to predict these processing requirements if a sufficient amount of training data is available. In particular, the processing requirement data may be continuous time-dependent data, which means that recurrent neural networks (RNNs), CNNs, and / or transformers may be very applicable in this setting, especially for accurate price prediction. In some embodiments, the system may use a model with time series forecasting and may use random forest algorithms, Bayesian RNNs, LSTMs, transformer-based models, CNNs, or other methods, or two or more combinations of these and the following: namely, neural ordinary differential equations (NODEs), stiff and non-stiff general ordinary differential equations (general ODEs), general stochastic differential equations (general SDEs), and / or general delayed differential equations (general DDEs).

[0054]

[0059] In some embodiments, the system may use models that include generative artificial intelligence. Generative artificial intelligence is a type of artificial intelligence technique that can generate various types of content, including text, images, audio, and synthetic data. Additionally or alternatively, the system may use LLMs of Small Language Models (SLMs). For example, an LLM is a type of artificial intelligence algorithm that uses deep learning techniques and very large datasets to understand, summarize, generate, and / or predict new content. The system may also use SLMs that are less energy-intensive and less costly by training smaller models by shifting the knowledge stored by the model from parameters to an external database, which not only reduces the number of parameters required but also makes it easier to update the model's knowledge. For example, instead of retraining the model, the system may simply expand the document database, which is done by feeding new data to the model and storing the resulting document embeddings.

[0055]

[0060] The system may receive user data via microservices and / or other means. For example, a microservice may comprise a collection of applications, each collecting one or more variables. For example, the system may extract user data from an API layer operating on the user's device or in a service provider (e.g., via a cloud service accessed by the user). Additionally or alternatively, the system may receive user data files (e.g., as real-time or near-real-time downloads and / or streaming).

[0056]

[0061] System 400 also includes an API layer 450. For example, in some embodiments, the system may be implemented as one or more APIs and / or API layers. In some embodiments, the API layer 450 may be implemented on a server 422 or a user terminal 424. Alternatively or additionally, the API layer 450 may reside on one or more cloud components 410. The API layer 450 (which may be a REST or web services API layer) may provide an isolated interface to the data and / or functionality of one or more applications. The API layer 450 may provide a common language-independent way of interacting with applications. Web services APIs provide a clearly defined contract called a WSDL, which describes the service in terms of the data types used to operate the service and exchange information. REST APIs typically do not have this contract. Instead, they are documented in client libraries for the most common languages ​​such as Ruby, Java, PHP, and JavaScript. Traditionally, SOAP web services have been adopted by enterprises to expose internal services or to exchange information with partners in B2B transactions.

[0057]

[0062] API Layer 450 can use various architectural configurations. For example, System 400 may be partially based on API Layer 450 and use resources such as a service repository and a developer portal, but with insufficient governance, standardization, and separation of concerns, while heavily employing SOAP and RESTful web services. Alternatively, System 400 may be fully based on API Layer 450, with separation of concerns between layers, services, and applications, as is the case with API Layer 450.

[0058]

[0063] In some embodiments, the system architecture can utilize a microservices approach. Such a system may use two types of layers: a front-end layer and a back-end layer where microservices reside. In this type of architecture, the role of the API layer 450 can be to provide integration between the front-end and back-end. In such cases, the API layer 450 may use a RESTful API (exposed to the front-end or even for communication between microservices). The API layer 450 may use AMQP (e.g., Kafka, RabbitMQ). The API layer 450 may adopt the early use of newer communication protocols such as gRPC and Thrift.

[0059]

[0064] In some embodiments, the system architecture can use an open API approach. In such cases, the API layer 450 can use commercial or open-source API platforms and their modules. The API layer 450 can use a developer portal. The API layer 450 can use strong security constraints that apply WAF and DDoS protection, and the API layer 450 can use RESTful APIs as a standard for external integration.

[0060]

[0065] Figure 5 shows a flowchart of the steps involved in generating dynamically updated metadata using a real-time artificial intelligence model, according to one or more embodiments. For example, the system may use process 500 (implemented, for example, on one or more of the system components described above) to generate dynamically updated metadata using a real-time artificial intelligence model (e.g., meta tags).

[0061]

[0066] In step 502, process 500 (for example, using one or more of the components described above) receives a first metadata tag requirement relating to the first metadata of the first media asset. For example, a system may receive a first metadata tag requirement relating to the first metadata of the first media asset. For example, a system may receive a first metadata tag requirement relating to the first metadata of the first media asset, the first metadata tag requirement indicating the right to use the first media asset (for example, the meta tag is used to indicate the right to use the digital asset and / or specifies the conditions for using the media asset, e.g., whether it can be used for commercial purposes, modified, or distributed). In another example, the metadata tag may identify information about the artificial intelligence technology used in the digital asset, which may include the type of artificial intelligence technology, the specific algorithm used, and / or other relevant information.

[0062]

[0067] In some embodiments, the system can select metadata tag requirements based on the type of content. For example, the system can determine a first type of a first media asset. Then, based on the first type, the system can determine first metadata tag requirements. For example, the system may use a taxonomy or controlled vocabulary, which is a predefined and standardized set of terms and categories for metadata tag requirements for a given media asset. Metadata tag requirements can be tagged and / or categorized using these taxonomies, thereby ensuring consistency and facilitating the selection of metadata tag requirements for a media asset type. Taxonomies may include hierarchies, relationships between terms, and rules of application. In some embodiments, the type may be determined manually. For example, a content creator, curator, or administrator may manually assign or enter metadata tag requirements based on their understanding of the media asset and its context. Alternatively or additionally, the system may use an automated process that analyzes the content of the media asset to extract relevant keywords, data, and / or concepts. For example, an NLP algorithm can identify key terms, entities, or phrases and suggest metadata based on their frequency, relevance, or semantic meaning. Additionally or alternatively, the system may use advanced techniques, such as machine learning algorithms, to analyze large datasets and identify patterns and correlations between content and metadata. These algorithms can learn from existing metadata associations and make predictions or recommendations for selecting appropriate metadata tag requirements for new media assets.

[0063]

[0068] In step 504, process 500 (for example, using one or more of the components described above) determines the metadata fields. For example, the system may determine a first metadata field relating to the first metadata based on the first metadata tag requirements. For example, the system may determine a first metadata field among several metadata fields of the first metadata based on the first metadata tag requirements. For example, in the license embodiment, the metadata fields are: <meta name=""使用権”content="クリエイティブコモンズ”"> This may include: In the above example of artificial intelligence, the system may include the name of a meta tag that reflects the type of artificial intelligence technology being used (for example, "

number

[0064]

[0069] In some embodiments, the system's determination of a first metadata field relating to a first metadata based on a first metadata tag requirement may include the system determining a second type of the first metadata tag requirement and determining the first metadata field based on the second type. For example, the system may select metadata fields based on the type of metadata tag requirement. For example, the system may use a taxonomy or controlled vocabulary, which is a predefined and standardized set of terms and categories for metadata fields based on the type of metadata tag requirement. Metadata fields can be tagged and / or categorized using these taxonomies, thereby ensuring consistency and facilitating the selection of metadata tag requirements for a given type of metadata tag requirement. Taxonomies may include hierarchies, relationships between terms, and rules of application. In some embodiments, the type may be determined manually. For example, a content creator, curator, or administrator may manually assign or enter metadata fields based on their understanding of the metadata tag requirement and context. Alternatively or additionally, the system may use an automated process that analyzes the content of the metadata tag requirement to extract relevant keywords, data, and / or concepts. For example, an NLP algorithm can identify important terms, entities, or phrases and suggest metadata fields based on their frequency, relevance, or semantic meaning. Additionally or alternatively, the system may use advanced techniques, such as machine learning algorithms, to analyze large datasets and identify patterns and correlations between metadata tag requirements and metadata fields. These algorithms can learn from existing metadata field-to-metadata tag requirement associations and make predictions or recommendations for selecting appropriate new metadata field-to-metadata tag requirement associations.

[0065]

[0070] In step 506, process 500 (for example, using one or more of the components described above) determines a content population function. For example, the system may determine a first content population function for a first metadata field, the first content population function including a self-contained code block that performs a rights verification task.

[0066]

[0071] In some embodiments, the first content population function may include a function that takes a specific type of content as input and outputs a different type of content. For example, the content population function may be used to update and / or modify content in one or more metadata fields. For example, the content population function may provide content for fields in meta tags. The content population function may populate metadata fields (e.g., by an artificial intelligence model) with content that may not exist at the time of media asset creation (e.g., similar and / or different images, similar text, similar audio, usage, usage location, use in other applications such as creating photos one after another using an artificial intelligence model). To generate this content, the content population function may specify a particular input type and output type. For example, the system may determine the input type of the first content population function. The system may then determine the output type of the first content population function.

[0067]

[0072] In step 508, process 500 (for example, using one or more of the components described above) generates metadata. For example, the system can generate first metadata using a first content population function, and the first metadata indicates a first usage right belonging to a first media asset. For example, in the artificial intelligence example above, the meta tag is the HTML code of the media asset (for example, <meta name=“

number

number

[0068]

[0073] In some embodiments, the first content population function may include a function that uses a specific algorithm to generate metadata. For example, the content population function may be used to update and / or modify content in one or more metadata fields and / or to generate new metadata using an algorithm. For example, determining the first content population function may include the system determining the algorithm of the first content population function and using the algorithm to determine the first metadata of the first media asset. For example, the content population function may generate content using one or more algorithms and techniques depending on the specific task and requirements. For example, the content population function may include a rule-based algorithm. These algorithms generate content using predefined rules and patterns. These algorithms often rely on templates, placeholders, and specific grammatical rules to create structured content. Additionally or alternatively, the content population function may include a Markov chain algorithm. A Markov chain is a probabilistic model that generates content based on the probability of transitions from one state to another. Markov chains can be trained on existing data, learn patterns, and generate new text sequences. Additionally or alternatively, content population functions may include RNN algorithms. RNNs are a type of deep learning algorithm commonly used for generating sequential data. They have a "memory" that allows them to capture context and generate consistent content. Variants of RNNs, such as Long Short-Term Memory (LSTM) and Gated Regressive Units (GRU), are commonly used for text generation tasks. Additionally or alternatively, content population functions may include Generative Adversarial Networks (GANs). GANs consist of two neural networks, a generator and a discriminator, that compete with each other.Generator networks learn to generate content (e.g., images, text) similar to the training data, while discriminator networks attempt to distinguish between real content and generated content. GANs have been successfully used for a variety of content generation tasks. Additionally or alternatively, content population functions may include transformer models. Transformer models such as OpenAI's GPT (Generatable Pre-trained Transformer) series, based on self-attention mechanisms, have made significant progress in NLP tasks, including content generation. These models learn from large amounts of text data and generate content that is consistent and contextually relevant. Additionally or alternatively, content population functions may include variational autoencoders (VAEs). VAEs are generative models that learn low-dimensional representations of input data. VAEs can be used to generate new content by sampling from the learned latent space and decoding it back into the original data domain.

[0069]

[0074] In some embodiments, the content population function can retrieve content that may be used to update the metadata of a media asset. For example, the system may continuously and / or periodically query one or more high-level public artificial intelligence systems, databases, and / or the internet to request updates to the underlying metadata of the media asset. The system may retrieve specific types of content. For example, generating first metadata using a first content population function may involve the system determining the content type of the first content population function and inputting the content type into the first metadata.

[0070]

[0075] For example, meta tags may include the ability to extract relevant data from synchronized rights and then map it to the appropriate metadata field. In some embodiments, the system identifies the relevant document, extracts relevant information, populates the metadata field, validates the current content, updates the content, and / or provides a unique identifier for the relevant metadata field.

number

[0071]

[0076] In some embodiments, the content population function can search for content that can be used to update the metadata of a media asset. For example, the system can sequentially query one or more high-level public artificial intelligence systems, databases, and / or the internet to request updates to the underlying metadata of a media asset. For example, a media asset can be extended to a digital token representing an NFT or physical asset by definition, thereby allowing the metatag to describe the underlying physical asset and change over time depending on the market value of such asset, publicly known attributes of such asset, etc. For example, the metatag of a digital token representing a car can change based on the car's mileage, car rating, car price on major sales platforms, car reputation, etc. For example, generating first metadata using a first content population function may include the system determining search criteria for the first content population function and performing a search based on the search criteria.

[0072]

[0077] In some embodiments, a content population function can update the metadata of a media asset based on a given operating frequency. For example, the system may use a clock signal to perform the operation at a given frequency. The clock signal provides a timing mechanism that synchronizes the operation of various components within the computer system. The system may also use other factors such as the efficiency of the processor architecture, cache size, memory speed, and / or software optimizations to determine when to execute the content population function. For example, generating first metadata using a first content population function may include the system determining the operating frequency of the first content population function and executing the first content population function based on that frequency.

[0073]

[0078] In some embodiments, a content population function can update the metadata of a media asset using a specific data source. For example, the system can continuously and / or periodically query one or more high-level public artificial intelligence systems, databases, and / or the internet to request updates to the underlying metadata of the media asset. For example, generating first metadata using a first content population function may include the system determining the data source for the first content population function and, based on the data source, determining the first metadata for the first media asset.

[0074]

[0079] In some embodiments, the system may parse a data source (e.g., a database, website, or the internet). For example, the system may establish a connection directly or via a network. This connection may be facilitated by an Internet Service Provider (ISP) or a local network. After establishing the connection, the system may send requests to a web server using a protocol such as the Hypertext Transfer Protocol (HTTP). These requests specify the required information, such as web pages, images, videos, or other resources. The web server receives the request and responds by sending back the requested web page. The server's response typically includes the requested content, along with metadata and other relevant information. The received web page may be written in Hypertext Markup Language (HTML), a language used to structure content on the web. The system can parse the HTML code using a computer's web browser or dedicated web crawling software to understand its structure and retrieve relevant data. The parsed HTML can be converted into a Document Object Model (DOM), which represents the structure of a web page as a tree-like data structure. This model allows the computer to access and manipulate various elements of a web page, such as headings, paragraphs, images, and links. When parsing a web page, a system may encounter links to other pages or resources. To effectively navigate the internet, computers follow these links, repeatedly submitting requests, retrieving web pages, and recursively parsing their content. For example, generating first metadata using a first content population function may involve the system parsing a data source using the first content population function and determining first metadata for a first media asset based on the parsing of the data source.

[0075]

[0080] In some embodiments, the system can extract data from third-party data sources. To extract data, the system may use automated metadata extraction. For example, the system may use text analysis. Text analysis involves analyzing the content of a file to extract metadata. For example, an NLP algorithm may be used to analyze the text of a document and extract keywords, topics, and named entities. This information can be used to automatically generate metadata tags for the file. Additionally or alternatively, the system may use image analysis. Image analysis involves analyzing the visual content of an image to extract metadata. For example, a computer vision algorithm may be used to identify objects, scenes, and other visual features in an image, which can then be used to automatically generate metadata tags for the image. Additionally or alternatively, the system may use audio and video analysis. Audio and video analysis involves analyzing the audio and visual content of a file to extract metadata. For example, a speech recognition algorithm may be used to transcribe the audio content of a video or audio file, which can then be used to automatically generate metadata tags for the file. For example, generating first metadata using a first content population function may include the system automatically extracting third-party data from a data source using a first content population function, where the system determines the first metadata for a first media asset based on third-party data.

[0076]

[0081] In some embodiments, the system can generate metadata for a given metadata tag name. For example, the metadata tag name may be used to indicate the type of artificial intelligence technology used in a media asset. The metadata tag name may specify the underlying technology or algorithm used to create or analyze the asset, providing valuable information to developers and researchers. For example, generating first metadata using a first content population function may involve the system determining the algorithm used by the first content population function and, based on the algorithm, determining the metadata tag name.

[0077]

[0082] In step 510, process 500 (for example, using one or more of the components described above) generates a media asset with metadata. For example, the system can generate a first media asset having first metadata (e.g., meta tags). In some embodiments, the system may reflect changes in the underlying media asset and its relationship to other media assets. For example, the metadata may provide a probability / confidence of how accurate the metadata is, whether (and / or to what extent) the media asset is similar to other media assets, and / or how (and / or by whom) the media asset was used. Furthermore, this solution overcomes the challenges with respect to metadata described above and does not introduce technical problems inherent in blockchain technology.

[0078]

[0083] In some embodiments, the system may continuously or periodically update the metadata of a media asset. For example, the system may generate second metadata using a first content population function. The system may then update the first media asset using the second metadata. For example, after generating a first media asset having the first metadata, the system may generate second metadata using a first content population function, where the second metadata indicates a second usage right belonging to the first media asset. The system may then update the first media asset using the second metadata.

[0079]

[0084] The steps or descriptions in Figure 5 are intended to be used in any other embodiments of this disclosure. In addition, the steps and descriptions described in relation to Figure 5 may be performed in an alternative order or in parallel to advance the objectives of this disclosure. For example, each of these steps may be performed in any order, in parallel or simultaneously to reduce lag or increase speed of the system or method. Furthermore, it should be noted that one or more of the steps in Figure 5 may be performed using any of the components, devices, or equipment discussed in relation to the above figures.

[0080]

[0085] Figure 6 shows a flowchart for using an artificial intelligence model to generate dynamically updated metadata in one or more embodiments. For example, the system may use specific algorithms and artificial intelligence models designed to enable automated / systematic optimization of the generation of dynamically updated metadata. For example, the system may use process 600 (implemented, for example, on one or more of the system components described above) to generate dynamically updated metadata (e.g., meta tags) using a real-time artificial intelligence model.

[0081]

[0086] In step 602, process 600 (for example, using one or more components as shown in Figure 4) determines the metadata tag requirements. For example, the system may receive a request for dynamically updated metadata. The request may further specify metadata tag requirements for the dynamically updated metadata.

[0082]

[0087] In step 604, process 600 (for example, using one or more components as shown in Figure 4) selects an artificial intelligence model based on metadata tag requirements. For example, the system may select an artificial intelligence model from multiple artificial intelligence models (for example, multiple artificial intelligence models as shown in Figure 3). The artificial intelligence model may use a Bayesian classifier, a decision tree learner, a decision rule classifier, a neural network, and / or a nearest neighbor algorithm.

[0083]

[0088] In step 606, process 600 (for example, using one or more components as shown in Figure 4) generates feature inputs for the selected artificial intelligence model. For example, the system can generate feature inputs having a format and / or values ​​that are normalized based on the model into which the feature inputs are input. For example, in some embodiments, the system can use latent representations (for example, as shown in Figure 3), in which a low-dimensional representation of the data may be used.

[0084]

[0089] In step 608, process 600 (for example, using one or more components as shown in Figure 4) inputs feature inputs into the artificial intelligence model. For example, the system may input feature inputs into the artificial intelligence model. For example, the system may generate a first feature input for the artificial intelligence model and input that input into the artificial intelligence model to generate dynamically updated metadata for the media asset.

[0085]

[0090] In step 610, process 600 (for example, using one or more components as shown in Figure 4) receives output. For example, the system may receive output from an artificial intelligence model. For example, the output may represent decisions used to generate metadata and / or create new meta tags. For example, each decision may include a decision on the content of a metadata field that may be based on one or more outputs from the artificial intelligence model.

[0086]

[0091] In step 612, process 600 (for example, using one or more components as shown in Figure 4) determines the metadata of the media asset based on the output. For example, the system may determine the content of one or more metadata fields based on the output from the artificial intelligence model.

[0087]

[0092] The steps or descriptions in Figure 6 are intended to be used in any other embodiments of this disclosure. In addition, the steps and descriptions described in relation to Figure 6 may be performed in an alternative order or in parallel to advance the objectives of this disclosure. For example, each of these steps may be performed in any order, in parallel, or substantially simultaneously to reduce system or method lag or increase speed. Furthermore, it should be noted that one or more of the steps in Figure 6 may be performed using any of the devices or equipment discussed in relation to Figures 1 to 5.

[0088]

[0093] The embodiments described herein are presented for illustrative purposes only, not limitation, and this disclosure is limited only by the following claims. Furthermore, it should be noted that features and limitations described in any one embodiment may apply to any other embodiment herein, and that flowcharts or examples relating to a certain embodiment may be combined with any other embodiment in an appropriate manner, performed in a different order, or performed in parallel. In addition, the systems and methods described herein may be performed in real time. It should also be noted that the systems and / or methods described herein may be applied to or used in accordance with other systems and / or methods.

[0089]

[0094] This technology will be better understood by referring to the embodiments listed below. 1. A method for generating dynamically updated metadata using a real-time artificial intelligence model. 2. The method of the preceding embodiment, further comprising: receiving a first metadata tag requirement relating to first metadata of a first media asset; determining a first metadata field relating to the first metadata based on the first metadata tag requirement; determining a first content population function for the first metadata field; generating first metadata using the first content population function; and generating a first media asset having the first metadata. 3. A method according to any one of the prior embodiments, further comprising: receiving a first metadata tag requirement relating to the first metadata of the first media asset; determining a first type of the first media asset; and determining the first metadata tag requirement based on the first type. 4. A method according to any one of the prior embodiments, further comprising determining a first metadata field relating to the first metadata based on a first metadata tag requirement, determining a second type of the first metadata tag requirement, and determining the first metadata field based on the second type. 5. A method according to any one of the prior embodiments, further comprising determining a first content population function relating to a first metadata field, determining the input type of the first content population function, and determining the output type of the first content population function. 6. A method according to any one of the prior embodiments, further comprising: generating first metadata using a first content population function; determining an algorithm for a first content population function; and determining first metadata for a first media asset using the algorithm. 7. A method according to any one of the prior embodiments, further comprising generating first metadata using a first content population function, determining the content type of the first content population function, and inputting the content type into the first metadata. 8. A method according to any one of the prior embodiments, further comprising generating first metadata using a first content population function, determining search criteria for the first content population function, and performing a search based on the search criteria. 9. A method according to any one of the prior embodiments, further comprising: generating first metadata using a first content population function; determining the frequency of operation of the first content population function; and executing the first content population function based on the frequency of operation. 10. A method according to any one of the prior embodiments, further comprising: generating first metadata using a first content population function; determining a data source for a first content population function; and determining first metadata for a first media asset based on the data source. 11. A method according to any one of the prior embodiments, further comprising generating first metadata using a first content population function, parsing a data source using a first content population function, and determining first metadata for a first media asset based on the parsing of the data source. 12. A method according to any one of the prior embodiments, further comprising generating first metadata using a first content population function, automatically extracting third-party data from a data source using a first content population function, and determining first metadata for a first media asset based on the third-party data. 13. A method according to any one of the prior embodiments, further comprising: generating first metadata using a first content population function; determining an algorithm used by the first content population function; and determining a metadata tag name based on the algorithm. 14. A method according to any one of the prior embodiments, further comprising generating second metadata using a first content population function and updating a first media asset using the second metadata. 15. One or more non-temporary computer-readable media that, when executed by a data processing device, store instructions causing the data processing device to perform the operations described in any of embodiments 1 to 14. 16. A system comprising one or more processors and a memory that stores instructions, when executed by the processors, causing the processors to perform the operations described in any of embodiments 1 to 14. 17. A system comprising means for carrying out any of embodiments 1 to 14.

Claims

1. One or more non-temporary computer-readable media containing instructions, wherein when the instructions are executed by one or more processors, Receiving the first metadata requirements for the first metadata of the first media asset, Based on the first metadata requirement, determine the first metadata field among a plurality of metadata fields relating to the first metadata, Determining a first content population function relating to the first metadata field, wherein the first content population function includes a self-contained code block that performs a verification task, The first metadata is generated using the first content population function, wherein the first metadata indicates first information belonging to the first media asset, To generate the first media asset having the first metadata, After generating the first media asset having the first metadata, generating the second metadata using the first content population function, wherein the second metadata indicates the second information belonging to the first media asset, Updating the first media asset using the second metadata, One or more non-temporary computer-readable media that cause an operation including the execution of such an operation.

2. A method for generating dynamically updated metadata using a real-time artificial intelligence model, Receiving the first metadata requirements for the first metadata of the first media asset, Based on the first metadata requirements, a first metadata field relating to the first metadata is determined, Determining a first content population function relating to the first metadata field, wherein the first content population function includes a self-contained code block that performs a verification task, The first metadata is generated using the first content population function, wherein the first metadata indicates first information belonging to the first media asset, To generate the first media asset having the first metadata, After generating the first media asset having the first metadata, generating the second metadata using the first content population function, wherein the second metadata indicates the second information belonging to the first media asset, Updating the first media asset using the second metadata, Methods that include...

3. Receiving the third metadata of the second media asset, The first metadata is generated using the third metadata and the first content population function, The method according to claim 2, further comprising:

4. Receiving user-generated metadata, Using the user-generated metadata and the first content population function, the first metadata is generated, The method according to claim 2, further comprising:

5. Receiving user profile data, Using the user profile data and the first content population function, the first metadata is generated, The method according to claim 2, further comprising:

6. Receiving a third metadata, wherein the third metadata includes a probability or confidence level relating to the second metadata, The second metadata is generated using the third metadata and the first content population function, The method according to claim 2, further comprising:

7. Receiving a third metadata, wherein the third metadata includes conditions used to provide a search function, The second metadata is generated using the third metadata and the first content population function, The method according to claim 2, further comprising:

8. Receiving a third metadata, wherein the third metadata includes the update period of the first content population function, The second metadata is generated using the third metadata and the first content population function, The method according to claim 2, further comprising:

9. Receiving a third metadata, wherein the third metadata includes a determination of an artificial intelligence model relating to the first content population function, The second metadata is generated using the artificial intelligence model, The method according to claim 2, further comprising:

10. Receiving a third metadata, wherein the third metadata includes a communication path for the first content population function, Updating the aforementioned first media asset using the aforementioned communication path, The method according to claim 2, further comprising:

11. Receiving the second metadata requirement regarding the third metadata of the second media asset, Based on the second metadata requirements, the second metadata field of the third metadata is determined, Determining a second content population function relating to the second metadata field, wherein the second content population function generates metadata based on metadata updates to the first media asset, The third metadata is generated using the second content population function, To generate the second media asset having the third metadata, After generating the second media asset having the third metadata, the fourth metadata is generated using the second content population function based on updating the first media asset using the second metadata, Updating the second media asset using the fourth metadata, The method according to claim 2, further comprising:

12. Receiving the first metadata requirement relating to the first metadata of the first media asset, Determining the first type of the first media asset, Based on the first type, the first metadata requirements are determined, The method according to claim 2, further comprising:

13. Determining the first content population function for the first metadata field is Determining the input type of the first content population function, Determining the output type of the first content population function, The method according to claim 2, further comprising:

14. The first metadata is generated using the first content population function. The algorithm for the first content population function is determined, Using the algorithm described above, the first metadata of the first media asset is determined, The method according to claim 2, further comprising:

15. The first metadata is generated using the first content population function. Determining the content type of the first content population function, Entering the content type into the first metadata, The method according to claim 2, further comprising:

16. The first metadata is generated using the first content population function. Determining the search criteria for the first content population function, Performing a search based on the aforementioned search criteria, The method according to claim 2, further comprising:

17. The first metadata is generated using the first content population function. The operation frequency of the first content population function is determined, Based on the aforementioned operating frequency, the first content population function is executed, The method according to claim 2, further comprising:

18. The first metadata is generated using the first content population function. Determining the data source for the first content population function, Based on the data source, the first metadata of the first media asset is determined, The method according to claim 2, further comprising:

19. The first metadata is generated using the first content population function. Analyzing the data source using the first content population function described above, Based on the analysis of the aforementioned data source, the first metadata of the first media asset is determined, The method according to claim 2, further comprising:

20. The first metadata is generated using the first content population function. Using the first content population function described above, third-party data is automatically extracted from the data source, Based on the aforementioned third-party data, the first metadata of the first media asset is determined, The method according to claim 2, further comprising:

Images