Intelligent accidental document discovery notification

By using prediction engines and prediction scenario models on a cloud-based content management platform, identifying documents that users may be interested in and providing suggestions cards, the problem of user positioning difficulties and waste of resources in large amounts of documents is solved, and rapid and resource efficiency is achieved document access.

CN119088765APending Publication Date: 2024-12-06GOOGLE LLC
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202411030645.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2018-07-23
Filing Date
2018-08-21
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

In a cloud-based content management platform, the process of users locating and retrieving documents of interest in a large number of documents becomes resource-intensive, and the content of the document changes over time, making it difficult for users to be sure whether the document is still interested, resulting in wasted network and processing resources.

Method used

Through the prediction engine, leverages machine learning models and predictive scenario models, identify documents that users may be interested in, and provides suggestions cards through graphical user interfaces (GUIs), reducing user search needs, providing smart buttons to quickly access documents, reducing networking and processing Resource consumption.

Benefits of technology

It significantly reduces the network resources and processing resources required to locate and retrieve documents from cloud storage, provides a wider range of documents than regular searches, improves users' reliable positioning of documents of interest, and reduces resource waste.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119088765A_ABST
    Figure CN119088765A_ABST
Patent Text Reader

Abstract

The invention relates to intelligent accidental document discovery notifications. A method of notifying a user of a cloud-based content management platform of documents includes identifying a first set of documents, wherein the first set of documents is hosted by the cloud-based content management platform and does not include one or more documents recently opened by the user; identifying one or more target documents from a first set of documents from the user based on an amount of topic overlap between the respective document and the currently working set of documents of the user, a number of viewing events for the respective document, and a number of collaboration events associated with the respective document, wherein the document set of the current work of the user comprises documents accessed by the user through the cloud-based content management platform within the last predetermined time period; and providing, for presentation to the user, a cloud-stored graphical user interface (GUI) of the user hosted by the cloud-based content management platform, the GUI identifying the one or more target documents.
Need to check novelty before this filing date? Find Prior Art

Description

Description of the case This application is a divisional application of a patent application with application number 201880096801.9 (PCT / US2018 / 047370), which was submitted to the China Patent Office on February 22, 2021, with an international application date of August 21, 2018, and the invention name is "Intelligent Unexpected Document Discovery Notification". Technical Field

[0001] Aspects and embodiments of the present disclosure relate to a cloud-based content management platform, and more particularly, to document discovery notifications from a cloud-based content management platform. Background Art

[0002] The cloud-based content management platform provides cloud storage for storing documents and editing applications (e.g., word processors, presentations, and spreadsheet applications) to collaboratively edit documents. Therefore, users of the cloud-based content management platform can collaborate over the network to create and store documents in cloud storage and edit documents, comment on documents, or share documents. The user will first access the documents associated with the user account via the main screen provided by the cloud-based content management platform. From the main screen, the user can locate a document in the document and access the located document. The cloud-based content management platform can additionally provide one or more applications, such as online calendars, emails, and messenger applications. Therefore, users can arrange meetings for projects and communicate with each other via emails or instant messages to promote online collaboration. Summary of the invention

[0003] The following is a brief summary of the disclosure in order to provide a basic understanding of some aspects of the disclosure. This summary is not an extensive overview of the disclosure. It is neither intended to identify the key or decisive elements of the disclosure, nor to characterize any scope of the specific embodiments of the disclosure or any scope of the claims. Its only purpose is to present some concepts of the disclosure in a simplified form as a preface to a more detailed description presented later.

[0004] In some embodiments, a system and method for notifying a user of a cloud-based content management platform of a change in the state of an electronic document associated with the user or a comment made in an electronic document associated with the user is disclosed. In an embodiment, a set of comments associated with documents accessed by the user via the cloud-based content management platform is determined. The set of comments includes one or more comments added to each document by other users within the last predetermined time period. One or more comments are selected from the set of comments to notify the user. The selection is based on the user's interaction with a comment topic associated with each comment from the set of comments and the characteristics of the comment topic. A comment topic is a group of multiple comments including a first comment and a second comment added in reply to the first comment. One or more documents associated with the one or more selected comments are determined. A graphical user interface (GUI) of the user's cloud storage hosted by the cloud-based content management platform is provided for presentation to the user. The GUI identifies the one or more documents. For each identified document, a corresponding selected comment is associated with the identified document.

[0005] When the GUI identifying the one or more electronic documents and corresponding selected comments is presented to the user, the user may provide one or more inputs related to one or more of the documents and / or comments identified in the GUI via the GUI, thereby enabling continued interaction between the user and the documents stored in the cloud-based content management platform.

[0006] In some other embodiments, a system and method for notifying a cloud-based content management platform of a document is disclosed. A first document set is identified. The first document set is hosted by the cloud-based content management platform and does not include one or more documents recently opened by the user. One or more target documents from the first document set of the user are identified based on the amount of overlap in topicality between the corresponding documents and the user's current working document set, the number of viewing events for the corresponding documents, and the number of collaboration events associated with the corresponding documents. The user's current working document set includes documents accessed by the user via the cloud-based content management platform within the last predetermined time period. A graphical user interface (GUI) of the user's cloud storage hosted by the cloud-based content management platform is provided for presentation to the user. The GUI identifies the one or more target documents.

[0007] The present disclosure provides a method for notifying a user of a cloud-based content management platform of a document. The method may include identifying, by a processor, a first document set, wherein the first document set is hosted by the cloud-based content management platform and does not include one or more documents recently opened by the user. The method may include identifying, by the processor, one or more target documents from the first document set of the user based on the amount of overlap in topicality between the corresponding document and the user's current working document set, the number of viewing events of the corresponding document, and the number of collaboration events associated with the corresponding document, wherein the user's current working document set includes documents accessed by the user via the cloud-based content management platform within a recent predetermined time period. The method may include providing, by the processor, a graphical user interface (GUI) of the user's cloud storage hosted by the cloud-based content management platform for presentation to the user, the GUI identifying the one or more target documents.

[0008] Identifying the first document set may include determining a second document set associated with one or more collaborators of the user based on recent events of one or more collaborators of the user and importance weights assigned to the one or more collaborators; determining a third document set related to documents in the document set currently being worked on by the user; generating the first document set by combining the second document set and the third document set and eliminating one or more documents recently opened by the user. Determining the second document set associated with one or more collaborators of the user based on recent events of one or more collaborators of the user and importance weights assigned to the one or more collaborators may include determining the one or more collaborators by identifying one or more other users of the cloud-based content management platform who provided at least one of a view event, an edit event, or a comment event to a document shared with the user in the cloud-based content management platform within a predetermined time period; determining the importance weight of each collaborator in a social network associated with the cloud-based content management platform based on the importance of the collaborator; and determining the second document set, the second document set including one or more documents having recent events of the one or more collaborators and importance weights assigned to the one or more collaborators. Determining the third document set related to the documents in the document set currently being worked on by the user may include: identifying key text of each document in the document set currently being worked on by the user; and determining the third document set including one or more documents having the key text. At least one of the second document set or the third document set may include one or more documents that have not yet been accessed by the user.

[0009] Generating the first document set may include at least one of the following: eliminating one or more documents that the user has not visited and have less than a predetermined number of one or more collaborators; and / or eliminating one or more documents that are not discoverable by a search performed via the cloud-based content management platform. Identifying the one or more target documents of the user may include at least one of the following: determining the number of viewing events associated with corresponding documents of other users or the one or more collaborators; and / or determining the number of collaboration events given the number of interaction events of the user with the corresponding document, the number of interaction events of the one or more other users with the corresponding document, and the likelihood that the user performs an interaction event of one or more other users involved in the corresponding document, wherein an interaction event is an event generated in response to an event of another user. Identifying the one or more target documents of the user may include: for each document in the first document set, applying as inputs to a machine learning model the amount of topical overlap between the corresponding document and the document set currently being worked on by the user, the number of viewing events of the corresponding document, and the number of collaboration events associated with the corresponding document; and obtaining an output from the machine learning model, the output indicating the probability that the user is interested in the corresponding document.

[0010] The GUI may include one or more suggestion cards. Each suggestion card may identify a corresponding or each target document. Each suggestion card may include at least one of: title information including a document type and title of the corresponding document; an image representation of the corresponding document; an action button providing access to the corresponding document; and / or reason text describing a reason for identifying the corresponding document.

[0011] When the GUI identifying the one or more electronic documents and corresponding selected comments is presented to the user, the user can provide one or more inputs related to one or more of the documents and / or comments identified in the GUI via the GUI. In this manner, the GUI can enable ongoing interaction between the user and the documents stored in the cloud-based content management platform.

[0012] The present disclosure provides a system, the system comprising a memory device and a processing device operably coupled to the memory device. The processing device may be configured to identify a first document set, wherein the first document set is hosted by the cloud-based content management platform and does not include one or more documents recently opened by the user. The processing device may be configured to identify one or more target documents from the first document set of the user based on the amount of topical overlap between the corresponding document and the user's current working document set, the number of viewing events of the corresponding document, and the number of collaboration events associated with the corresponding document, wherein the user's current working document set includes documents accessed by the user via the cloud-based content management platform within a recent predetermined time period. The processing device may be configured to present to the user a graphical user interface (GUI) of the user's cloud storage hosted by the cloud-based content management platform, the GUI identifying the one or more target documents.

[0013] The processing device can be configured such that, after identifying the first document set, the processing device is configured to: determine a second document set associated with one or more collaborators of the user based on recent events of the one or more collaborators and importance weights assigned to the one or more collaborators; determine a third document set related to documents in the document set currently being worked on by the user; and generate the first document set by combining the second document set and the third document set and eliminating one or more documents recently opened by the user.

[0014] The processing device may be configured to determine the second document set associated with one or more collaborators of the user based on recent events of the one or more collaborators and importance weights assigned to the one or more collaborators. The processing device may be configured to determine the one or more collaborators by identifying one or more other users of the cloud-based content management platform who provided at least one of a view event, an edit event, or a comment event to a document shared with the user in the cloud-based content management platform within a predetermined time period; determine the importance weight of each collaborator based on the importance of the corresponding collaborator in a social network associated with the cloud-based content management platform; and determine the second document set, which includes one or more documents having recent events of the one or more collaborators and importance weights assigned to the one or more collaborators.

[0015] The processing device may be configured to determine the third document set related to the documents in the document set currently being worked on by the user, the processing device may be configured to identify the key text of each document in the document set currently being worked on by the user; and determine the third document set including one or more documents having the key text. The processing device may be configured to identify the one or more target documents of the user, the processing device may be configured to perform at least one of the following: determine the number of viewing events associated with the corresponding documents of other users or the one or more collaborators; or determine the number of collaboration events in view of the number of interaction events of the user with the corresponding document, the number of interaction events of the one or more other users with the corresponding document, and the likelihood of the user performing an interaction event of one or more other users involved in the corresponding document, wherein an interaction event is an event generated in response to an event of another user.

[0016] The GUI may include one or more suggestion cards, each of which may identify each or a corresponding target document. Each suggestion card may include at least one of: title information including a document type and title of a corresponding document; an image representation of the corresponding document; an action button providing access to the corresponding document; and / or reason text describing a reason for identifying the corresponding document.

[0017] The present disclosure provides a non-transitory computer-readable medium storing instructions which, when executed, cause a processing device to: identify a first document set, wherein the first document set is hosted by the cloud-based content management platform and does not include one or more documents recently opened by the user; identify one or more target documents from the first document set of the user based on an amount of topical overlap between the corresponding documents and the user's currently working document set, a number of viewing events for the corresponding documents, and a number of collaboration events associated with the corresponding documents, wherein the user's currently working document set includes documents accessed by the user via the cloud-based content management platform within a last predetermined time period; and provide a graphical user interface (GUI) of the user's cloud storage hosted by the cloud-based content management platform for presentation to the user, the GUI identifying the one or more target documents.

[0018] The instructions, when executed to identify the first document set, may cause the processing device to: determine a second document set associated with one or more collaborators of the user based on recent events of the one or more collaborators and weight importances assigned to the one or more collaborators; determine a third document set related to documents in the document set currently being worked on by the user; and generate the first document set by combining the second document set and the third document set and eliminating one or more documents recently opened by the user.

[0019] The instructions, when executed, may cause the processing device to: determine the one or more collaborators by identifying one or more other users of the cloud-based content management platform who have provided at least one of a view event, an edit event, or a comment event to a document shared with the user in the cloud-based content management platform within a predetermined time period; determine the importance weight of each collaborator based on the importance of the collaborator in a social network associated with the cloud-based content management platform; and determine the second document set, which includes one or more documents having the one or more collaborators' recent events and the importance weights assigned to the one or more collaborators.

[0020] When the instruction is executed, it determines the third document set related to the documents in the document set currently being worked on by the user, causing the processing device to: identify the key text of each document in the document set currently being worked on by the user; and determine the third document set including one or more documents having the key text.

[0021] The GUI may include one or more suggestion cards, each suggestion card identifying a corresponding target document or each target document. Each suggestion card may include at least one of: title information including a document type and title of the corresponding document; and an image representation of the corresponding document; an action button providing access to the corresponding document; and / or reason text describing a reason for identifying the corresponding document. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Aspects and embodiments of the present disclosure will be more fully understood from the detailed description given below and from the drawings of the aspects and embodiments of the present disclosure, however, the present disclosure should not be limited to specific aspects or embodiments but is only for explanation and understanding.

[0023] Figure 1 An example of a system architecture that implements some aspects of the present disclosure is illustrated.

[0024] Figure 2 Illustrated are example user interfaces for suggestion cards that notify a user of an electronic document and / or an event or other state change of an electronic document.

[0025] Figure 3 A flow chart illustrating aspects of a method for selecting documents to notify a user of a cloud-based content management platform according to one embodiment of the present disclosure.

[0026] Figure 4 A flow chart depicting aspects of a method for determining documents associated with selected one or more reviews according to one implementation of the present disclosure.

[0027] Figure 5 A flow chart depicting aspects of a method for predicting documents for users of a cloud-based content management platform according to another embodiment of the present disclosure.

[0028] Figure 6 A flow chart depicting aspects of a method for predicting documents for a user based on multiple predictive scenario models according to one embodiment of the present disclosure.

[0029] Figure 7 A flow chart illustrating aspects of a method for selecting documents for suggestion according to one embodiment of the present disclosure.

[0030] Figure 8 A block diagram depicts an example computing system operating in accordance with one or more aspects of the present disclosure. DETAILED DESCRIPTION

[0031] The cloud-based content management platform can provide cloud storage for electronic documents and editing applications such as word processors, presentations, and spreadsheet applications, among others. A user can log in to a cloud-based content management platform and be presented with a main screen of a cloud-based content management platform. The main screen used herein refers to a graphical user interface (GUI) that lists documents stored in cloud storage associated with a user account. Cloud storage used herein refers to a repository of online electronic documents associated with a user account. Such documents may include documents owned or created by a user and / or one or more groups of which the user is a member, and / or documents shared with a user and / or one or more user groups. A user can select a document from the main screen and access the document using an editing application provided by a cloud-based content management platform. Among others, accessing a document as used herein refers to opening, viewing, editing a document, or adding comments to a document, etc. Among others, comments used herein refer to posts added to a specific portion of a document, posts added in association with a suggested edit on a portion of a document, or posts assigning tasks to a user in association with a specific portion of a document.

[0032] As more users collaborate via cloud-based content management platforms, the number of documents stored in cloud storage has grown exponentially. As a result, users may need to spend a lot of time locating and retrieving documents of interest to the user from the main screen of the cloud-based content management platform. In addition, the user may even forget the documents that the user has previously opened or needed, or the user may not be fully aware of all the documents available to the user. This is particularly serious in collaborative work systems, where multiple users can potentially create, edit, share, or comment on documents, and the number of documents of the cloud-based content management platform may be so large that the user is only aware of a small percentage of the documents available.

[0033] Aspects and embodiments of the present disclosure solve the above-mentioned defects, in particular, by predicting documents that may be of interest to the user among documents associated with a user account. Documents associated with a user account refer to documents that can be accessed and / or viewed by the user. The prediction technology of the present disclosure can predict documents containing comment topics that may interest the user due to comments recently added to the comment topic. The prediction technology may take into account the user's interaction with the comment topic and the characteristics of the comment topic. Moreover, the prediction technology of the present disclosure can predict unexpected documents. As used herein, an unexpected document refers to a document that the user has not recently visited or viewed (i.e., the user may not know (e.g., because the user has not visited the document before)) or no longer remembers such a document (e.g., because it has been longer than a determined or predetermined time period since the user last interacted with the document) and the user may be interested in. An unexpected document may also be referred to as a target document. In order to predict a target or unexpected document, the prediction technology may consider the level of relevance of the document to the document that the user is working on, the popularity of the document among other users, and / or the level of collaboration involving the document. Among other things, heuristic methods or machine learning methods can be used to implement the prediction technology.

[0034] Aspects and embodiments of the present disclosure also address the above-mentioned deficiencies by notifying the user of a predicted document using a GUI component that explains the reason for the notification and provides quick access to the document. The notification technology of the present disclosure may produce a suggestion card (a GUI component that presents information about the corresponding document) for each predicted document. The notification technology may involve extracting the reason for the prediction and inserting the reason into the suggestion card. In addition, the notification technology may provide the user with a user interface component, such as a smart button, for quick access to the document.

[0035] Thus, aspects and embodiments of the present disclosure provide users with quick access to documents that the user may be interested in. As a result, the need for time-consuming searches is eliminated, thereby reducing processing and network resources used for searches.

[0036] A technical problem that may be solved by various aspects and embodiments of the present disclosure is that the continued growth in the number of documents accessible to users via cloud-based content management platforms over computer networks has made the process of locating and retrieving documents more resource intensive. For example, locating a particular document in cloud storage may involve a user manually reviewing a large corpus of documents until the desired document is found. This may result in an inefficient use of network resources and processing resources. For example, network resources may be consumed by transferring each document in the corpus from cloud storage to a user device, whereupon processing resources may be consumed by opening each document at the user device to allow manual review by the user. This may be a particular problem in situations where a mobile user device has limited network resources and / or processing resources.

[0037] Another technical problem is that the content of a document is not fixed, but can change over time, so that even if a user is aware of the existence of a particular document, they cannot be sure what the document contains and whether the content of the document may be of interest to the user. Even if a user has previously accessed a document, they do not know whether the current content of the document is the same as when they last accessed the document. This means that the user may have to retrieve and access the document in order to see whether the content of the document has been modified since they last accessed the document and whether the user is still interested in the current content, which again requires the use of network resources and processing resources (which is wasted if, for example, there have been no significant changes in the document since the user last accessed the document).

[0038] To address the technical issues identified above, aspects and embodiments of the present disclosure may notify a user about documents that may be of interest to the user without requiring explicit action by the user. This may significantly reduce the network resources and processing resources required to locate and retrieve documents from cloud storage. Moreover, this may provide notifications to users about the status of electronic documents stored, for example, in cloud storage and / or about changes in status. Alternatively or additionally, aspects and embodiments of the present disclosure may provide a limited number of notifications (i.e., suggestion cards), thereby reducing the number of network resources and processing resources required when sending notifications. Alternatively or additionally, aspects and embodiments of the present disclosure may allow users to more easily and reliably locate specific documents of interest, thereby significantly reducing the network bandwidth and processing resources required when retrieving documents from cloud-based storage.

[0039] Aspects and embodiments of the present disclosure may notify a user of information about a document only to the extent that the document is predicted to be of interest to the user. This may reduce the resources required to send notifications, and may result in a more user-friendly or less intrusive user interface, at least as compared to a situation in which the user is notified of every document event.

[0040] Therefore, aspects and embodiments of the present disclosure reduce or conserve computing resources and network bandwidth required for document searching while providing users with a wider range of documents than conventional search engines can provide.

[0041] Figure 1 An example of a system architecture 100 that implements some aspects of the present disclosure is illustrated.

[0042] The system architecture 100 includes a cloud-based environment 110 connected to user devices 120A-120Z via a network 130. The cloud-based environment 110 refers to a collection of physical machines that host applications that provide one or more services (e.g., content management) to multiple user devices 120A-120Z via the network 130. The network 130 can be a public network (e.g., the Internet), a private network (e.g., a local area network (LAN) or a wide area network (WAN)), or a combination thereof. The network 130 can include a wireless infrastructure that can be provided by one or more wireless communication systems such as wireless fidelity (WiFi) hotspots connected to the network 130 and / or wireless carrier systems that can be implemented using various data processing equipment, communication towers, etc. Additionally or alternatively, the network 130 can include a wired infrastructure (e.g., Ethernet).

[0043] The cloud-based environment 110 may include a server 112 and a data store 114. The data store 114 may be separate from the server 112 and communicatively coupled to the server 112, or the data store 114 may be part of the server 112. In some embodiments, the data store 114 may be included as part of the user devices 120A-120Z. The data store 114 may include cloud storage for storing documents for each user. The data store 114 may store metadata associated with the documents stored in the cloud storage. Such metadata may include an access control list for the user (e.g., who the creator or owner is, who can view and edit, who has shared the document with), and events associated with the document. An event as used herein refers to a user activity (e.g., an open event, a view event, an edit event, or a comment event) caused by accessing a document. For each event, the data store 114 may store metadata associated with the event, such as the type of event, the creator of the event, the content, and a timestamp. For comment events, the data store 114 may additionally store information about the associated comment topic. A comment topic is a set of multiple comments having a first comment and a second comment added as a reply to the first comment. Some examples of such information include the number of comments in a thread, the user associated with each comment in the comment thread, the status of the comment thread (e.g., open, closed, if open, how long it has been open), and the relationship between the comments in the comment thread (e.g., which comment was the initial comment that started the comment thread and which comment was a reply to that comment). The data store 114 may also store the predictive scenario model and its output, as explained in detail below with respect to the prediction engine 116.

[0044] The server 112 may include a physical machine (e.g., a server machine, a desktop computer, etc.) that includes one or more processing devices that are communicatively coupled to a memory device and an input / output (I / O) device. The server 112 may host a cloud-based content management platform 115 and may include one or more processing devices. In an alternative embodiment, the cloud-based content management platform 115 may be hosted by user devices 120A-120Z (e.g., as a standalone application) and operate as a peer in a peer-to-peer environment. In other alternative embodiments, applications on user devices 120A-120Z may interact with the cloud-based content management platform 115 and may perform some of the functionality of the content management platform 115.

[0045] The cloud-based content management platform 115 may enable the user to access documents and view the home screen user interface (UI) of the cloud-based content management platform as corresponding UIs 124A-124Z. Additionally, in some embodiments, the cloud-based content management platform 115 may provide UIs 124A-124Z that include suggestion cards 200 to notify the user of documents that the user is interested in. The cloud-based content management platform 115 may further provide suggestion cards 200 to enable the user to quickly access documents or perform actions on documents without opening the documents.

[0046] In an embodiment, the UI 124A-124Z of the cloud-based content management platform 115 can be a web page rendered by a web browser and displayed in a web browser window on the user device 120A-120Z. In another embodiment, the UI 124A-124Z can be provided by a mobile application or a desktop application. For example, the UI 124A-124Z can be displayed by a native application executed on the user device 120A-120Z. The native application can be separated from the web browser.

[0047] The server 112 includes a prediction engine 116 to predict documents to be recommended to the user and notify the user of these documents. In an embodiment, the prediction engine 116 may use a prediction scenario model to determine the documents to be recommended to the user. The prediction engine 116 may use a heuristic method or a machine learning method to implement the prediction scenario model, such as Figures 3 to 5 Discussed in detail.

[0048] A predictive scenario model may be a model for predicting notified documents or events that a user may be interested in. The prediction engine 116 may utilize multiple predictive scenario models when predicting which document or event should be notified to the user. Some examples of predictive scenario models may be related to regular, targeted, or unexpected document discovery and / or comments. Each predictive scenario model may be customized to a specific scenario. For example, a predictive scenario model based on regular document discovery may be designed to predict documents in a scenario where a user routinely accesses documents via a cloud-based content management platform 115. In this case, the prediction engine 116 may obtain output from a predictive scenario model based on regular document discovery, with documents that are frequently accessed and recently accessed by the user having a higher probability than other documents that are infrequently accessed and recently accessed by the user. A predictive scenario model based on targeted or unexpected discovery may operate in a scenario where the user is not aware of or has forgotten about documents that may be helpful to the user (i.e., targeted or unexpected documents) because there are too many documents available via the cloud-based content management platform 115. Documents that may be helpful to the user may include documents that are related to the document that the user is currently working on, that the user's collaborators have recently worked on, and that are popular among other users. The following is about Figure 5The use of the prediction scenario model by the prediction engine 116 based on target or unexpected findings is further described.

[0049] In addition, the comment-based predictive scenario model can operate in scenarios where users interact with comments from a specific comment topic (e.g., a comment topic in which the user is actively involved (e.g., a comment topic created by the user, or the user has subscribed to a reply), a comment topic in which many users are involved). Figure 3 to Figure 4 The use of this predictive scenario model based on comments by the prediction engine 116 is further described. In addition, there may be a predictive scenario model for a scenario in which a user responds to an email, chat, or calendar event that mentions the user or is from other users with whom the user frequently interacts. In such a scenario, the predictive scenario model may rank calendar events for meetings with other users with whom the user frequently interacts higher than other calendar events open to ordinary users of the cloud-based content management platform 115.

[0050] The prediction engine 116 can provide input data to the prediction scenario model and obtain output data from the model. The input data can vary among the prediction scenario models, such as Figures 3 to 5 As will be discussed in more detail below, the input to the predictive scenario model can be data from various signals associated with the document that represent the specific scenario on which the model is based. The output data of the predictive scenario model can be a predicted scenario score that represents the probability that the user will be interested in the document based on the corresponding scenario. If recommended, the predicted scenario score can also represent how the user is likely to interact with the document. In the case of using multiple predictive scenario models, the prediction engine 116 can obtain a predicted scenario score from each model and, based on the data from the prediction engine 116 relative to the model, the predicted scenario score can be obtained. Figure 7 The prediction scores of the different types of predictive scenario models described in detail determine the overall document score. The prediction engine 116 can then use the overall document score to determine one or more documents for notification.

[0051] In some implementations, the prediction engine 116 can generate a suggestion card for each document for which a notification is selected based on the overall document score and / or the prediction score (e.g., see Figure 2 ). The suggestion card may include reason text 209 describing the reason for notifying the user of the document and one or more smart buttons 210 (210A, 210B) that enable the user to quickly access the document.

[0052] The user devices 120A-120Z may include one or more processing devices communicatively coupled to a memory device and an I / O device. The user devices 120A-120Z may be a desktop computer, a laptop computer, a tablet computer, a mobile phone (e.g., a smart phone), or any suitable computing device. As discussed above, the user devices 120A-120Z may each include a web browser and / or a client application (e.g., a mobile application or a desktop application). A user may view a suggestion card 200 or a home screen of a cloud storage including a suggestion card 200 via a web browser or a client application using the cloud-based content management platform 115. For example, once a user logs into the cloud storage, the suggestion card 200 may be presented to the user for notification on the home screen. The user may select a suggestion card 200 to access the document represented by the suggestion card 200, or select a smart button (e.g., Figure 2 210A or 210B) to respond to events associated with a document directly from the suggestion card 200 without opening the document. In this way, the user device 120A associated with the user can send a request to display the suggestion card 200 and transmit the user selection. In general, where appropriate, functions described as being performed by the cloud-based content management platform 115 and / or prediction engine 116 in one embodiment can also be performed on user devices 120A to 120Z in other embodiments. In addition, functionality attributed to a particular component may be performed by different or multiple components operating together.

[0053] Figure 2 An example of a user interface presenting a suggestion card 200 to notify a user of an event and / or other state change of an electronic document and / or an electronic document stored in the user's cloud storage is illustrated. The cloud storage is hosted by a cloud-based content management platform 115. In one embodiment, the suggestion card 200 may be presented to the user as part of a home screen of the cloud-based content management platform. Such a home screen may be provided to the user after the user successfully logs into the cloud storage of the cloud-based content management platform 115.

[0054] In some embodiments, the suggestion card 200 may take the form of a card (i.e., a box similar to a card), such as Figure 2. Suggestion card 200 represents a document and an event associated with the document. The prediction engine 116 can construct suggestion card 200 to notify the user about suggestions of documents that are predicted to be of interest to the user. Suggestion card 200 can allow the user to quickly access the documents represented in suggestion card 200. In other embodiments, the prediction engine 116 can compose suggestion card 200 to notify the user and enable the user to respond to the event represented by suggestion card 200. When the prediction engine 116 uses one prediction scenario model, the prediction engine 116 can select the document with the highest prediction scenario score to generate suggestion card 200. In the case where the prediction engine 116 uses multiple prediction scenario models, the prediction engine 116 can determine which documents to notify the user based on the overall document score, using, for example, information about Figure 7 The method shown and described in determines the overall document score, which incorporates all available prediction scenario scores for the document. The prediction engine 116 can limit the number of suggestion cards 200 to be presented to the user. In an embodiment, the prediction engine 116 can constrain the number and position of suggestion cards 200 based on a particular type of predictive scenario model. For example, the prediction engine 116 can obtain at most one document from the predictive scenario model based on a goal or unexpected discovery, and place such documents only after the third position. The prediction engine 116 can also incorporate temporal diversity (e.g., showing that some documents are to be opened now, while some documents are to be opened in the next few hours when the suggestion card is presented). Refer to the following Figures 3 to 7 A detailed description of various embodiments of a method for selecting documents to suggest or notify a user is described. Figure 2 As illustrated in , suggestion card 200 includes multiple sections, such as document information section 202 , document thumbnail section 204 , and event section 206 .

[0055] The document information section 202 may include information about the document, such as the title of the document, the type of the document (e.g., word, spreadsheet, presentation slide), and overflow menu options (e.g., a vertically oriented oval). Such document information may be represented as an icon or text. When the user selects an overflow menu option (such as an option for sharing the document, adding the document to favorites, locating the document in the user's cloud storage, and removing the suggestion card 200), the prediction engine 116 may present more options. The document thumbnail section 204 may include thumbnails of specific portions of the document. The prediction engine 116 may select a portion associated with the event of the suggestion card 200. For example, when the event is a comment event, the prediction engine 116 may take a fragment of a paragraph to which the comment event is added. In another embodiment, in the absence of any reference to a portion of the document, the thumbnail may be an image of the first page of the document or a default image showing the type of document and the title of the document. In some embodiments, the prediction engine 116 may open the document in response to the user's selection of the document thumbnail section 204.

[0056] The event section 206 may include details of the event. In some implementations, the prediction engine 116 may add information about other users associated with the event (e.g., name and profile picture), the content of the event, and reason text describing the reason for notifying the user of the event and / or document. For example, in Figure 2 , the event portion 206 includes a profile picture and the name "Ben DuPree" who created the comment event and the content of the comment event - "This is looking good + Simon Smith, could you take a look at this when you have a minute?". Below the name "Ben DuPree", the event portion 206 further includes a reason text ("Suggested because you created the comment thread of this comment") 209 and the timestamp of the comment event.

[0057] In an embodiment, the prediction engine 116 may determine the event and reason text based on the data used as input to the prediction scenario model, which data contributes most to the output of the model (i.e., the prediction scenario score). For example, the prediction engine 116 may use a prediction scenario model based on a comment that uses data specifying whether a user has created a comment topic associated with the comment event. The prediction engine 116 may then compose the event portion 206 based on information from the comment. In addition, if the prediction engine 116 determines that the score for the criterion is the highest among the other criteria, the prediction engine 116 may compose reason text, such as "Suggested because you created the comment thread of this comment" to explain the reason for notifying the user about the comment event. In another example, the prediction engine 116 may use a prediction scenario model based on a goal or unexpected discovery that utilizes data specifying the topicality of the document and the amount of overlap in the document set that the user is currently working on. Topical overlap can represent a measure or similarity score of how related the documents are to the document set that the user is currently working with, and can be determined, for example, based on a predefined set of criteria (such as overlap level, occurrence of words or phrases identified as related to a common field or topic in a stored database, Jaccard similarity, cosine distance, Euclidean distance, etc.).

[0058] If the prediction engine 116 determines that the score from the criterion is the highest score among the other criteria, the prediction engine 116 may compose a reason text such as "Suggested based on documents for Project X". In another embodiment, the prediction engine 116 may compose the reason text based on the latest event of the document. For example, if the latest event associated with the document is a calendar event to which the document has been attached, the prediction engine 116 may use "12:30pm today" as the calendar text for the scheduled meeting time as the reason text.

[0059] The prediction engine 116 may determine the reason text based on the prediction scenario model that caused the document to be selected for notification. In an embodiment, when the prediction engine 116 uses multiple prediction scenario models and the document has a reason text from Figure 7 When a document has a high overall document score and the contribution from a specific prediction scenario model to the overall document score is the largest, the prediction engine 116 may mention the description of the specific prediction scenario model in the reason text such as “Suggested based on a serendipitous discovery”.

[0060] The event section 206 may further include one or more smart buttons 210 (210A, 210B) to enable a user to quickly access a document or respond to an event suggested to the user via the suggestion card 200. In an embodiment, the prediction engine 116 may include an open button (enabling a user to open a document) or a preview button (enabling a user to preview a document without opening the document using an application of the cloud-based content management platform 115) in the event section 206. The prediction engine 116 may also enable a user to open a document of the suggestion card 200 in response to detecting a user selection (e.g., double-click) of any portion of the suggestion card 200. In another embodiment, the prediction engine 116 may select one or more smart buttons 210 (210A, 210B) based on the type of event included in the suggestion card 200. For example, if the event notified to the user via the suggestion card 200 is a comment event, the prediction engine 116 may select the reply smart button 210A or the resolve smart button 210B. The prediction engine 116 may enable a user to directly reply to the comment event within the suggestion card 200 without opening the document via the reply smart button 210A. In an embodiment, in response to the prediction engine 116 detecting the user's selection of the reply smart button 210A, the prediction engine 116 may provide a text box in the suggestion card 200 for the user to type the content of the reply and provide an add button to add the reply content to the corresponding comment thread.

[0061] Additionally, the prediction engine 116 may enable the user to close a comment thread from a suggestion card 200 without requiring the user to open the document. In another example, if an email or chat event is notified to a user via a suggestion card 200, the prediction engine 116 may include a reply button and a locate item button. The locate item button may enable the user to find the email or chat message of which the user is being notified by presenting an inbox with the email highlighted or a chat window with the chat message highlighted. Thus, the prediction engine 116 may generate any other smart button that provides a shortcut for the user to respond to the event described in the event portion 206. In some embodiments, the prediction engine 116 may limit the number of smart buttons 210 of a particular type (210A, 210B) (e.g., at most one calendar event action may be included in the last set of suggestion cards to be presented).

[0062] Figure 3 A flow chart illustrating aspects of a method 300 for predicting or selecting documents for notifying a user of a cloud-based content management platform 115 according to one embodiment of the present disclosure.

[0063] Method 300 may be performed by Figure 1The method 300 may be performed by the prediction engine 116 of the user device 120. In another embodiment, the method 300 may be performed by a client application executed by one or more processing devices of the user device 120. The method 300 may be performed for each user of the cloud-based content management platform 115. In addition, the method 300 may be performed when the user requests the server 112 to access the user's cloud storage or home screen via the cloud-based content management platform 115. In some embodiments, the method 300 may be performed to notify the user of comments made in a document associated with the user via the cloud-based content management platform 115 by using a comment-based predictive scenario model.

[0064] For simplicity of explanation, method is depicted and described as a series of actions in this article. However, actions according to the present disclosure may occur in various orders and / or simultaneously, and other actions are not presented and described in this article at the same time. In addition, it may not be required that all actions shown are implemented according to the method of the disclosed subject matter. In addition, it should be understood and appreciated by those skilled in the art that these methods can be alternatively represented as a series of related states via state diagrams or events. Additionally, it should be understood that the methods disclosed in this specification can be stored on articles to facilitate the conveying and transfer of such methods to computing devices. The term "article" as used herein is intended to encompass a computer program accessible from any computer-readable device or storage medium.

[0065] At block 302, the prediction engine 116 may determine a set of comments associated with documents accessed by the user via the cloud-based content management platform 115. The prediction engine 116 may identify documents that have been shared with the user or created by the user as documents that the user has access to. To identify such documents, the prediction engine 116 may query the access control list of the document, which lists the user as a creator, viewer, editor, or subscriber (i.e., a user shared with the document). Based on the identified documents, the prediction engine 116 may determine the comments added to each document by other users. For each comment, the prediction engine 116 may determine information related to the corresponding document, such as a timestamp and the user who created the corresponding comment. The prediction engine 116 may then compile a set of comments that includes comments added by other users to documents accessed by the user within a recent predetermined time period (i.e., recent comments). For example, the user may be working on multiple projects using the cloud-based content management platform 115 and collaborating via documents. In this case, when the user logs into the cloud storage via the cloud-based content management platform 115, the prediction engine 116 may determine the documents that the user has accessed (e.g., the documents that the user is working on). The prediction engine 116 can then identify the comments that the collaborators made to the document, and among those comments, determine the comments that were made within the last two days. In this way, the prediction engine 116 can determine the candidates for comments that the user may be interested in responding to.

[0066] At block 304, the prediction engine 116 may select one or more comments from the set of comments determined at block 302 to notify the user based on information about the review topic to which the respective review under consideration belongs. A review topic is a group of multiple comments having a first comment and a second comment added as a reply to the first comment. The prediction engine 116 may select comments from the set based on the user's interaction with the review topic associated with each comment. In one embodiment, the prediction engine 116 may identify the review topic associated with the respective comment for each comment in the set of comments from block 302. The prediction engine 116 may then determine the user's interaction with the review topic in order to select the comment to be notified to the user. For example, in some embodiments, the prediction engine 116 may determine whether the user has added an initial comment (i.e., a comment that starts a review topic) in the review topic. The prediction engine 116 may also determine whether the user has added any comments to the review topic as a reply to the initial comment in the review topic or any other comment in the review topic.

[0067] At block 304, the prediction engine 116 may further perform the selection of comments for notification based on characteristics of the comment topic associated with each comment. Characteristics of the comment topic may be measured in various ways. Some examples of characteristics of the comment topic are the state of the comment topic (whether the comment topic is open or closed), the number of comments included in the comment topic, the number of different users involved in the comment topic, the proportion of comments in the comment topic that are added or closed by a user, the proportion of comments in the comment topic that are added by other users, and the length of a period of time that the comment topic remains open.

[0068] The prediction engine 116 may further perform selecting comments for notification at 304 based on importance weights assigned to each user (i.e., any other user) who added or generated the comment. The prediction engine 116 may determine the importance weights based on the importance of the corresponding user in a social network associated with the cloud-based content management platform 115. Some examples of social networks may be, among others, an organizational chart of users of the cloud-based content management platform 115 (i.e., higher user weights associated with higher positions in the chart), a collaborator chart based on frequency and recency from user response history (i.e., higher weights for more frequent collaborators), an affinity chart that includes affinity scores (e.g., a similarity measure based on affinity) of other users (i.e., higher weights for users with higher affinity scores), and a reputation chart (i.e., higher weights for more popular users).

[0069] When selecting a comment at box 304, the prediction engine 116 may determine a comment score for each comment in the set from box 302 based on the user's interaction with the comment topic and the characteristics of the comment topic. In another embodiment, the prediction engine 116 may further incorporate an importance weight for each user who adds the corresponding comment. The comment score may indicate the probability that the user is interested in the corresponding comment or the document containing the corresponding comment. The prediction engine 116 may select one or more comments for notification based on the comment score. The prediction engine 116 may determine the comment score based on the comment using a prediction scenario model. The prediction engine 116 may use a heuristic method or a machine learning method to construct a prediction scenario model. In the heuristic method, the prediction engine 116 may apply a weighted sum to the prediction scenario model and further normalize the weighted sum to between 0 and 1. The prediction engine 116 may assign a weight and a value between 0 and 1 to each criterion, which represents the user's interaction with the comment topic and the characteristics of the comment topic. The prediction engine 116 may update the assigned weights and values ​​to better reflect the user's interests.

[0070] In the machine learning method, the prediction engine 116 can train a machine learning model. Hereinafter, the prediction scenario model based on comments may also be referred to as a machine learning model. The prediction engine 116 may include a training engine capable of training a prediction scenario model. The prediction scenario model may be a machine learning model corresponding to a model artifact created by the training engine using training data, the training data including training inputs and corresponding target outputs (i.e., the correct answer for the corresponding training input). In some embodiments, the prediction engine 116 may generate training data, which includes one or more training inputs and one or more target outputs. For comments added by other users to documents involving the user, the training input may include data indicating the user's interaction with the comment topic and the characteristics of the comment topic. In another embodiment, the target input may additionally include data indicating the importance weights of other users who added comments. The target output may include responses, such as replies to comments, closing the comment topic to which the corresponding comment belongs, accepting or rejecting suggested edits associated with comments, marking the to-do tasks assigned in the comments as completed by the user, or no response. In another embodiment, the training data may not be specific to a particular user, but to various users of the cloud-based content management platform 115.

[0071] The training engine can find patterns in the training data that map the training inputs to the target outputs (i.e., the answers to be predicted) and provide a machine learning model that captures these patterns. The machine learning model can consist of, for example, a single level of linear or nonlinear operations (e.g., a support vector machine [SVM]), or can be a deep network (i.e., a machine learning model consisting of multiple levels of nonlinear operations). An example of a deep network is a neural network with one or more hidden layers, and such a machine learning model can be trained by, for example, adjusting the weights of the neural network according to a back propagation learning algorithm, etc. For convenience, although some embodiments may employ SVM or other types of learning machines instead of or in addition to neural networks, the remainder of this disclosure will also refer to the implementation as a neural network. Once the prediction engine 116 determines that the machine learning model is ready for use, the prediction engine 116 can provide data for the user's interaction with the comment topic and the characteristics of the comment topic as input to the trained machine learning model (i.e., a prediction scenario model). The prediction engine 116 can then obtain an output from the trained machine learning model that indicates the likelihood that the user will be interested in a particular comment from the comment topic, for example, as a comment score with a numerical value between 0 and 1.

[0072] In some implementations, at block 304, the prediction engine 116 may further consider information about documents associated with the review, the review itself, and user activity when determining the review score. Some examples of information about documents associated with each review in the collection from block 302 may include the number of users involved in the corresponding document, the number of users with different types of access to the corresponding document, the recency of events in the document, the recency of a particular type of event (e.g., edit, view, or comment), the proportion of review topics in the document that were started or closed by a user. For information about the review, the prediction engine 116 may further consider the length of the review text and whether the review is directed to the user by explicitly mentioning the user in the review text.

[0073] After obtaining the review scores, the prediction engine 116 may rank each review in the review set from block 302 based on the review scores. In one embodiment, the prediction engine 116 may assign a higher review score with a higher ranking to a review than a review with a lower review score. The prediction engine 116 may then select a subset of reviews from the review set based on the rankings.

[0074] At block 306, the prediction engine 116 may determine one or more documents associated with the one or more selected reviews from block 304. Figure 4 The implementation of determining the document in block 306 is described in detail.

[0075] At box 308, the prediction engine 116 can utilize the cloud-based content management platform 115 to provide a GUI for presentation to the user based on the user account. The user account can define which documents stored on the cloud-based content management platform 115 are associated with the user (the user has access to). The storage of such documents is referred to as the user's cloud storage in this article. The prediction engine 116 can provide the GUI when the user logs into the cloud-based content management platform 115. In another embodiment, the prediction engine 116 can provide the GUI as part of the main screen of the cloud-based content management platform. The GUI can identify the documents associated with the selected comments. For each document identified in the GUI, the GUI can further identify the corresponding selected comments associated with the corresponding document. If multiple selected comments are associated with a document, the prediction engine 116 can select the most recent comment or the comment with the highest ranking from the multiple comments. In another embodiment, the prediction engine 116 can determine the comment with the highest ranking from the comment topic with the highest ranking among other comment topics in the document. The following will be with respect to Figure 4 Describes in detail the ranking of the review topic.

[0076] In some embodiments, the GUI may include Figure 2 One or more suggestion cards 200 to identify documents and comments. The suggestion card can identify the document and the comments associated with the document, as described above with respect to Figure 2 For example, the suggestion card 200 may include title information of the document, such as the document type and title of the document in the document information section 202, an image representation (e.g., a thumbnail) of the document in the document thumbnail section 204, and one or more smart buttons 210 (210A, 210B) for replying to a comment from the suggestion card without opening the document, another one or more smart buttons 210 (210A, 210B) for closing a comment from the suggestion card without opening the document, text for the content of the comment, and text describing the reason for selecting the corresponding comment in the event section 206.

[0077] Thus, when a GUI identifying one or more electronic documents and corresponding selected comments is presented to the user, the user may provide one or more inputs related to one or more of the documents and / or comments identified in the GUI via the GUI (e.g., via smart buttons 210A, 210B). This enables interaction between the user and documents stored in cloud storage via the cloud-based content management platform 115.

[0078] Figure 4 Describes an embodiment of the present disclosure for use in Figure 3 The block 306 determines the Figure 3Flowchart of aspects of method 400 for selecting a document associated with a comment selected at block 304 of the method 400. Method 400 may be performed using a comment-based predictive context model.

[0079] At block 402, the prediction engine 116 may identify Figure 3 The prediction engine 116 may query the data store 114 for at least one document set associated with the review selected at block 304 of FIG. Figure 3 Any commented documents selected at box 304 .

[0080] Once the documents have been identified, at block 404, the prediction engine 116 may determine a document score for each document based on the review scores of the reviews selected from block 304. The document score may indicate a probability that a user may be interested in interacting with the document. In an embodiment, the prediction engine 116 may assign the highest document score to the document with the review with the highest review score according to the prediction scenario model. In another example where the document is associated with multiple reviews selected from block 304 and the multiple reviews may belong to different review topics in the document, the prediction engine 116 may identify a set of review topics from each document, wherein each review topic is associated with at least one of the selected reviews from block 304. The prediction engine 116 may then determine a review topic score for each review topic based on the review scores of the selected reviews associated with the review topic. The review topic score may indicate a probability that a user is interested in the review topic. In one embodiment, the prediction engine 116 may use a weighted sum based on the reviews as a prediction scenario model to calculate the review topic score. In another embodiment, the prediction engine 116 may assume each review score as an independent event according to the prediction scenario model based on the reviews, and determine the probability of the independent event as the review topic score. Once the prediction engine 116 determines the review topic scores for the review topics in each document, the prediction engine 116 may determine a document score for each document based on a weighted sum of the review topic scores as a model of the review-based prediction scheme.

[0081] At block 406, the prediction engine 116 may rank each document in the document set identified at block 402 based on the document score. At block 408, the prediction engine 116 may select one or more documents from the document set of block 402 based on the ranking. Figure 3 At block 308 , the prediction engine 116 may present the selected document to the user for notification via the suggestion card 200 .

[0082] Figure 5 A flow chart depicting aspects of a method 500 for predicting documents for users of a cloud-based content management platform 115 according to another embodiment of the present disclosure.

[0083] Method 500 may be performed by Figure 1 In another embodiment, the method 500 may be performed by a client application executed by one or more processing devices of the user device 120. The method 500 may be performed for each user of the cloud-based content management platform 115. In addition, the method 500 may be performed when the user requests the server 112 to access the user's cloud storage or home screen via the cloud-based content management platform 115. In some embodiments, the method 500 may be performed to notify the user of the cloud-based content management platform 115 of the document suggested via the cloud-based content management platform 115 using a predictive scenario model based on a goal or serendipity. In one embodiment, the method 500 may be performed simultaneously with the method 300.

[0084] At block 502, the prediction engine 116 may identify a first set of documents. In some embodiments, the documents in the first set are hosted by a cloud-based content management platform 115. In addition, the first set of documents may exclude documents that the user has recently opened. Therefore, the prediction engine 116 may select documents of the cloud-based content management platform 115 that the user may no longer remember or may not have opened at all as candidates for suggestion. For example, the documents that the user may no longer remember may include, for example, at least one of the following: documents that have expired from a cache or history or may not have taken any action (such as being opened, read, edited, modified, or otherwise interacted with by the user within a determined or predetermined time period).

[0085] In some embodiments, the prediction engine 116 may determine the first set of documents based on events (i.e., collaboration levels) from collaborators of the cloud-based content management platform 115. For example, the prediction engine 116 may identify a set of documents associated with the collaborator, or documents that the collaborator may open, view, or edit. The prediction engine 116 may further determine from the set of documents documents that have recent events (e.g., open, view, edit, or comment events) by the collaborator as candidate documents for suggestion. In an embodiment, the prediction engine 116 may define any user of the cloud-based content management platform 115 that has worked on the same document as a collaborator. For example, the prediction engine 116 may determine a collaborator by identifying another user of the cloud-based content management platform 115 that has performed a view, edit, or comment event with respect to a document shared with the user via the cloud-based content management platform 115 within a predetermined time period. In one embodiment, the prediction engine 116 may determine the first set of documents based on the number of views (i.e., popularity levels) of the collaborator or any other user. In some embodiments, the prediction engine 116 may further identify candidate documents for suggestion based on importance weights assigned to each collaborator, such as with respect to Figure 3 Detailed description of block 304 of FIG.

[0086] When determining the first document set as a candidate for suggestion, the prediction engine 116 may further incorporate documents related to the document into the document set of the user's current work (i.e., the level of relevance). The prediction engine 116 may determine the document set of the user's current work. The document set of the user's current work may include documents accessed by the user via the cloud-based content management platform 115 within the last predetermined time period. For example, any document that the user has opened, viewed, or edited via the cloud-based content management platform 115 in the past week may be considered as the document set of the user's current work. In another embodiment, the prediction engine 116 may define the document set of the user's current work using only a specific type of access (e.g., editing). Once the prediction engine 116 identifies the document set of the user's current work, the prediction engine 116 may determine the key text of each document in the set of the user's current work. The key text may be indexed and represent the topic of the document content. The key text may be extracted from the title or content of the corresponding document. The key text may be a word, a phrase, or a sentence. The key text may be manually typed by the user associated with the document, or may be automatically generated via the cloud-based content management platform 115. The key text may be used to query the associated documents stored in the data store 114. Prediction engine 116 may query data store 114 for any documents associated with key text of the user's currently working document set. In this way, prediction engine 116 may determine document sets related to documents in the user's currently working collection as candidates for suggestion.

[0087] In some embodiments, the document set associated with the collaborator and another document set related to the collection of the user's current work may include documents that the user has not yet accessed. In this way, the prediction engine 116 can select documents that the user is not aware of because the user has never accessed those documents as candidates for suggestion. In an embodiment, the prediction engine 116 can filter out the documents that the user has recently opened from these two collections in order to select documents that the user may no longer remember as candidates for suggestion. Therefore, the prediction engine 116 can generate a merged document set (i.e., a first document set) for suggestion by combining the document set associated with the collaborator and the document set related to the collection of the user's current work and filtering out the documents that the user has recently opened. In an embodiment, the prediction engine 116 can further filter out documents that may not make sense to the user because less than a threshold number of collaborators are working on the document. Therefore, the prediction engine 116 can exclude documents that the user has not accessed and have events performed by less than a predetermined number of collaborators from the first set. In another embodiment, the prediction engine 116 can also eliminate documents that are set to be undiscoverable by searches performed via the cloud-based content management platform 115 (e.g., via tags in the metadata of the document).

[0088] At block 504, the prediction engine 116 may identify one or more target documents from the first document set to suggest to the user. The target or unexpected documents may be documents that the user may not be aware of (e.g., because the user may not have opened, edited, or otherwise interacted with them) or no longer remember (e.g., because it is longer than a determined or predetermined period of time since the user last interacted with the document), and that may be of interest to the user. In one embodiment, the prediction engine 116 may identify the target documents based on at least one of a level of relevance of the corresponding document from the set of block 502 to the user's currently working document set, a level of popularity of the corresponding document among users, a level of collaboration associated with the corresponding document, and the like.

[0089] The prediction engine 116 may measure the level of relevance as the amount of overlap in topicality between the corresponding document and the user's current working document set. The overlap in topicality may represent a measure or similarity score of how the document is related to the user's current working document set, and may be determined, for example, based on a predefined set of criteria such as overlap level, occurrence of words or phrases identified in a stored database related to a common field or theme, Jaccard similarity, cosine distance, Euclidean distance, etc. The prediction engine 116 may identify the topicality of a document from, among other things, key text assigned to the document or specific terms or document content repeated in the document. The prediction engine 116 may then measure how similar the topicality of the corresponding document is to the documents in the user's current working collection.

[0090] The prediction engine 116 can determine the popularity level of a document based on the number of viewing events within a predetermined time period. The viewing events can be conducted by any user, or by users in the user's social network (e.g., collaborators, friends, or group members) via the cloud-based content management platform 115.

[0091] The prediction engine 116 can measure the collaboration level of a document based on the number of collaboration events associated with the corresponding document. A collaboration event is an interaction event (i.e., an event generated in response to an event of another user) that is exchanged back and forth between users in the context of using a document for collaboration. Collaboration events can occur in the form of comment threads (e.g., posting a comment and replying to the comment), suggesting an edit (e.g., suggesting an edit and accepting or rejecting the suggestion), or sharing a document (e.g., sharing a document and opening the document). Therefore, in one embodiment, the prediction engine 116 can determine the number of collaboration events in view of the number of interaction events of the user, the number of interaction events of other users, and the likelihood of the user performing an interaction event with other users involved in the corresponding document. The prediction engine 116 can estimate the likelihood of a user performing an interaction event with others based on the user's past response history.

[0092] In some embodiments, the prediction engine 116 may determine a target score or an unexpected score for each document in the first set from box 502 based on at least one of the relevance level, popularity level, or collaboration level. The prediction engine 116 may then use the target score or the unexpected score to select a document with a high score as a target document to suggest to the user. The target score or the unexpected score may indicate the probability that the user is interested in the corresponding document when the user is not aware of or has forgotten the corresponding document that may be helpful to the user. The prediction engine 116 may adopt a heuristic method or a machine learning method when determining the target score or the unexpected score. In the heuristic method, the prediction engine 116 may apply a weighted sum as a prediction scenario model, and further normalize the weighted sum to be between 0 and 1. The prediction engine 116 may assign a weight and a value between 0 and 1 to each relevance level, popularity level, and collaboration level.

[0093] In the machine learning method, the prediction engine 116 can train the machine learning model. In some embodiments, the prediction engine 116 can generate training data, which includes one or more training inputs and one or more target outputs. Hereinafter, the prediction scenario model based on the target or unexpected discovery may also be referred to as a machine learning model. The training input may include data indicating the relevance level, popularity level and / or collaboration level of each document in the first set from box 502. The target output may include a user's response, such as opening, viewing, editing or sharing a response document, or no response. In another embodiment, the training data may not be specific to a specific user, but to various users of the cloud-based content management platform 115. Once the prediction engine 116 determines that the machine learning model is ready for use, the prediction engine 116 can provide the current input to the trained machine learning model (e.g., current data for the relevance level, popularity level and collaboration level of the documents in the first set from block 502). Then, the prediction engine 116 can obtain the output from the trained machine learning model as a target score or an unexpected score with a numerical value between 0 and 1. In one embodiment, the prediction engine 116 may apply, for each document in the first set from block 502, the amount of topical overlap between the corresponding document and the user's current working document set, the number of viewing events for the corresponding document, and / or the number of collaboration events associated with the corresponding document as input to the trained machine learning model. The prediction engine 116 may then obtain an output from the trained machine learning model. The output may indicate the likelihood that the user is interested in the corresponding document.

[0094] At box 506, the prediction engine 116 may provide a GUI 200 of the user's cloud storage hosted by the cloud-based content management platform 115 for presentation to the user. The GUI 200 may identify one or more target documents. In some embodiments, the GUI 200 may include one or more suggestion cards. Each suggestion card may identify each target document. In addition, each suggestion card may include title information in the document information section 202 (e.g., the type and title of the corresponding document), an image representation of the corresponding document in the document thumbnail section 204 (e.g., a thumbnail), and one or more smart buttons 210 (210A, 210B) that provide access to the corresponding document (e.g., open, view, edit) and reason text that describes the reason for identifying the corresponding document in the event section 206, as described above with respect to Figure 2 Described in detail.

[0095] Figure 6 A flow chart depicting aspects of a method 600 for predicting documents for a user based on multiple predictive scenario models according to one embodiment of the present disclosure.

[0096] The method 600 may be performed by the prediction engine 116. In another embodiment, the method 600 may be performed by a client application executed by one or more processing devices of the user device 120. The method 600 may be performed when the user requests the server 112 to access the cloud storage or the home screen via the cloud-based content management platform 115. The method 600 may be performed when the prediction engine 116 uses multiple prediction scenario models to predict documents for recommendation to the user. Examples of prediction scenario models include: Figure 2 The prediction scenario models of conventional, targeted or unexpected findings and comments are described. In an embodiment, in response to receiving a request from a user to access the cloud storage or home screen of the cloud-based content management platform, the prediction engine 116 can run multiple prediction scenario models in real time. In another embodiment, the prediction engine 116 can periodically run all or some of the prediction scenario models and store the prediction scores in the data storage 114.

[0097] The prediction engine 116 may obtain an output (e.g., a prediction scenario score) from each prediction scenario model. In an embodiment where the prediction engine 116 utilizes two prediction scenario models, at box 602, the prediction engine 116 may determine a first prediction scenario score for a first document set from a first prediction scenario model. In some embodiments, the prediction engine 116 may identify a document set associated with the input data required for the first prediction scenario model as a first document set. The prediction engine 116 may run the first prediction scenario model by providing the required information associated with each document from the first set as input, and obtain a prediction score for each document. The prediction scenario score may represent the probability that a user is interested in the corresponding document under the corresponding scenario of the prediction scenario model. Such a prediction score may be in the range of 0 to 1. In an embodiment, the prediction engine 116 may run a comment-based prediction scenario model by providing data representing the user's interaction with the comment topic in the document and the characteristics of the comment topic of each document as input. The prediction engine 116 may then obtain the prediction scenario score for the document.

[0098] At box 604, the prediction engine 116 may use the second prediction scenario model to determine a second prediction scenario score for the second document set. The prediction engine 116 may run a prediction scenario model based on a target or unexpected discovery by providing data representing the level of relevance, collaboration, and popularity of each document as input. The prediction engine 116 may then obtain the prediction scenario score for the document. In some embodiments, the document set used in one prediction scenario model may be different from the document set of another model because different models may be based on different scenarios. For example, when running a prediction scenario model based on comments, the prediction engine 116 may use data from documents with comments recently added by other users; however, for a prediction scenario model based on a target or unexpected discovery, the prediction engine 116 may use data from documents with comments recently added by other users, but exclude those recently opened by the user. In another example, the document set used in the prediction scenario model based on a target or unexpected discovery and the document set based in the conventional model may be mutually exclusive because the former model is constructed under a scenario that explicitly excludes documents recently opened by the user; while the latter model considers documents recently opened or accessed by the user in any other way.

[0099] At block 606, the prediction engine 116 may select documents from the first document set and the second document set for recommendation based on the respective first prediction scenario scores and the second prediction scenario scores. Figure 7 Implementations of block 606 are discussed in detail. In some implementations, the prediction engine 116 may further provide a plurality of suggestion cards for the document selected from block 606 on the home screen GUI.

[0100] Figure 7 FIG. 1 is a diagram showing a method for performing a Figure 6 Flowchart of aspects of method 700 of selecting a document for suggestion at block 606 .

[0101] At block 702, the prediction engine 116 may Figure 6 The prediction engine 116 may identify the weight of each scenario used in the multiple predictive scenario models in blocks 602 and 604. At block 704, the prediction engine 116 may also identify any attenuation factors for each scenario. In some embodiments, the value of the attenuation factor may be initially pre-set by the prediction engine 116. For example, the prediction engine may initially set the attenuation factor for the target or unexpected discovery scenario with a small value (e.g., 0.01). When calculating the overall document score in block 708, the attenuation factor may be used to exponentially reduce the contribution of the predicted scenario score from a particular predictive scenario model. In some embodiments, the prediction engine 116 may update the weight and attenuation factor for each scenario. At block 706, the prediction engine 116 may select the attenuation factor from the prediction engine 116. Figure 6Blocks 602 and 604 of the method determine a prediction scenario score assigned to each document in the first and second document sets. Each prediction scenario score is associated with a different prediction scenario model. In an embodiment, some documents may have only one prediction score assigned, while other documents may have more than one prediction score.

[0102] At block 708, the prediction engine 116 may calculate an overall document score for each document based on the predicted scenario scores and the weights and decay factors corresponding to each scenario. The prediction engine 116 may add the product of the predicted scenario scores, weights, and any decay factors for each scenario to determine the overall document score. In embodiments, the prediction engine 116 may further consider the standard deviation of the different predicted scenario scores when calculating the overall document score.

[0103] At block 710, the prediction engine 116 may rank the documents based on the overall document score. In an embodiment, the prediction engine 116 may rank the documents in ascending or descending order of the overall document score. The prediction engine 116 may select a limited number of documents for recommendation based on the ranking at block 712.

[0104] Figure 8 800 is a block diagram of an example computing system that operates in accordance with one or more aspects of the present disclosure. In various illustrative examples, computer system 800 may be associated with Figure 1 The computer system 800 corresponds to any one of the computing devices within the system architecture 100. In one embodiment, the computer system 800 may be the server 112. In another embodiment, the computer system 800 may be each of the user devices 120A-120Z.

[0105] In some embodiments, computer system 800 can be connected to other computer systems (for example, via a network such as a local area network (LAN), an intranet, an extranet, or the Internet). Computer system 800 can operate with the ability of a server or client computer in a client-server environment, or operate as a peer-to-peer computer in a peer-to-peer or distributed network environment. Computer system 800 can be provided by a personal computer (PC), a tablet PC, a set-top box (STB), a personal digital assistant (PDA), a cellular phone, a web device, a server, a network router, a switch or a bridge or any device that can (sequentially or otherwise) perform a group of instructions specifying the action to be taken by the device. In addition, the term "computer" should include executing a group (or groups) of instructions individually or jointly to perform any one or more of the computer's any set of methods described herein.

[0106] In another aspect, the computer system 800 may include a processing device 802, a volatile memory 804 (e.g., random access memory (RAM)), a non-volatile memory 806 (e.g., read only memory (ROM) or electrically erasable programmable ROM (EEPROM)), and a data storage device 816 that may communicate with each other via a bus 808.

[0107] The processing device 802 may be provided by one or more processors such as a general purpose processor (such as, for example, a complex instruction set computing (CISC) microprocessor, a reduced instruction set computing (RISC) microprocessor, a very long instruction word (VLIW) microprocessor, a microprocessor that implements other types of instruction sets, or a microprocessor that implements a combination of various types of instruction sets) or a special purpose processor (such as, for example, an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), a digital signal processor (DSP), or a network processor).

[0108] The computer system 800 may also include a network interface device 822. The computer system 800 may also include a video display unit 810 (eg, LCD), an alphanumeric input device 812 (eg, keyboard), a cursor control device 814 (eg, mouse), and a signal generating device 820.

[0109] The data storage device 816 may include a non-transitory computer-readable storage medium 824 on which may be stored instructions 826 encoding any one or more of the methods or functions described herein, including instructions for implementing Figure 1 The prediction engine 116 is provided with instructions for implementing the methods described herein.

[0110] The instructions 826 may also reside, completely or partially, within the volatile memory 804 and / or within the processing device 802 during execution thereof by the computer system 800, and thus, the volatile memory 804 and the processing device 802 may also constitute machine-readable storage media.

[0111] Although the computer-readable storage medium 824 is shown as a single medium in the illustrative example, the term "computer-readable storage medium" shall include a single medium or multiple media (e.g., a centralized or distributed database, and / or associated caches and servers) that store one or more sets of executable instructions. The term "computer-readable storage medium" shall also include any tangible medium that can store or encode a set of instructions executed by a computer, the instructions causing the computer to perform any one or more of the methodologies described herein. The term "computer-readable storage medium" shall include, but is not limited to, solid-state memories, optical media, and magnetic media.

[0112] In the above description, many details are set forth. However, it will be apparent to those of ordinary skill in the art who benefit from the present disclosure that the present disclosure can be practiced without these specific details. In some cases, well-known structures and devices are shown in block diagram form rather than in detail in order to avoid obscuring the present disclosure.

[0113] Some portions of the detailed description have been presented in terms of algorithms and symbolic representations of operations on data bits within a computer memory. These algorithmic descriptions and representations are the means used by those skilled in the art of data processing to most effectively communicate the essence of their work to other persons skilled in the art. An algorithm is here and generally considered to be a self-consistent sequence of steps leading to a desired result. These steps are steps requiring physical manipulation of physical quantities. Typically, although not necessarily, these quantities take the form of electrical or magnetic signals that can be stored, transferred, combined, compared, and otherwise manipulated. For general reasons, it has proven convenient to sometimes refer to these signals as bits, values, elements, symbols, characters, terms, numbers, etc.

[0114] It should be remembered, however, that all of these and similar terms will be associated with the appropriate physical quantities and are merely convenient labels applied to these quantities. Unless expressly stated otherwise, as will be apparent from the following discussion, it should be understood that, throughout the specification, discussions utilizing terms such as "receive," "display," "rank," "identify," "determine," "notify," "predict," "suggest," "provide," "present," and the like refer to the actions and processes of a computer system or similar electronic computing device that manipulates and transforms data represented as physical (e.g., electronic) quantities within the computer system's registers and memories into other data similarly represented as physical quantities within the computer system's memories or registers or other such information storage, transmission, or display devices.

[0115] For simplicity of explanation, method is depicted and described as a series of actions in this article. However, actions according to the present disclosure may occur in various orders and / or simultaneously, and other actions are not presented and described in this article at the same time. In addition, it may not be required that all actions shown are implemented according to the method of the disclosed subject matter. In addition, it should be understood and appreciated by those skilled in the art that these methods can be alternatively represented as a series of related states via state diagrams or events. Additionally, it should be understood that the methods disclosed in this specification can be stored on articles to facilitate the conveying and transfer of such methods to computing devices. The term "article" as used herein is intended to encompass a computer program accessible from any computer-readable device or storage medium.

[0116] Certain embodiments of the present disclosure also relate to a device for performing the operations herein. This device can be constructed for a predetermined purpose, or it can include a general-purpose computer selectively activated or reconfigured by a computer program stored in the computer. Such a computer program can be stored in a computer-readable storage medium, such as but not limited to any type of disk including a floppy disk, an optical disk, a CD-ROM, and a magneto-optical disk, a read-only memory (ROM), a random access memory (RAM), an EPROM, an EEPROM, a magnetic card or an optical card, or any type of medium suitable for storing electronic instructions.

[0117] References throughout this specification to "one embodiment" or "an embodiment" mean that a particular feature, structure, or characteristic described in conjunction with the embodiment is included in at least one embodiment. Therefore, the phrases "in one embodiment" or "in an embodiment" appearing in various places throughout this specification do not necessarily all refer to the same embodiment. In addition, the term "or" is intended to mean an inclusive "or" rather than an exclusive "or". In addition, the word "example" or "exemplary" is used herein to mean used as an example, instance, or illustration. Any aspect or design described herein as "exemplary" is not necessarily to be construed as being preferred or advantageous over other aspects or designs. On the contrary, the use of the word "example" or "exemplary" is intended to present the concept in a concrete way.

[0118] It should be understood that the above description is intended to be illustrative, rather than restrictive. After reading and understanding the above description, many other embodiments will be apparent to those skilled in the art. Therefore, the scope of the present disclosure should be determined with reference to the attached claims together with the full scope of equivalents given by such claims.

[0119] In addition to the above description, users may be provided with controls that allow users to select whether and when the system, program, or feature described herein can enable the collection of user information (e.g., information about the user's social network, social actions or activities, profession, user's preferences, or the user's current location) and whether the user is sent content or communication from the server. In addition, certain data may be processed in one or more ways before it is stored or used so that personally identifiable information is removed. For example, the user's identity may be processed so that personally identifiable information cannot be determined for the user, or the user's geographic location may be generalized (such as to a city, zip code, or state level) when location information is obtained so that the user's specific location cannot be determined. Therefore, users may have control over what information is collected about the user, how the information is used, and what information is provided to the user.

Claims

1. A method for notifying a user of a cloud-based content management platform of a document, the method comprising: In response to a login request from the user to log into the cloud-based content management platform, one or more target documents that may be of interest to the user and have not been opened by the user or have been opened by the user for more than a predetermined period of time are selected from a plurality of documents associated with the user account of the user on the cloud-based content management platform, wherein selecting the one or more target documents comprises: identifying, by a server of the cloud-based content management platform, a first set of unexpected documents among the plurality of documents associated with the user account of the user, wherein the first set of unexpected documents associated with the user account of the user includes documents uploaded to the cloud-based content management platform by the user and documents uploaded to the cloud-based content management platform by other users, wherein each unexpected document in the first set is a document that has expired from cache or has been opened by the user for more than the predetermined time period; and The one or more target documents of the first unexpected document set from the user are identified by the server of the cloud-based content management platform based on: (i) an amount of topical overlap between unexpected documents from the first unexpected document set that have expired from cache or have been last opened by the user for more than the predetermined time period and a document set of the user's work that was last opened within the predetermined time period, (ii) a number of view events for the unexpected documents, and (iii) a number of collaboration events associated with the unexpected documents; and In response to the login request of the user, a home screen graphical user interface (GUI) of the cloud-based content management platform is provided by the server of the cloud-based content management platform for presentation to the user, the home screen GUI identifying the one or more target documents.

2. The method according to claim 1, wherein: Identifying the first unexpected document set includes: determining a second set of documents associated with the one or more collaborators based on events of the one or more collaborators of the user and importance weights assigned to the one or more collaborators; determining a third document set related to the documents in the document set of the user's work; and The first unexpected document set is generated by combining the second document set and the third document set and eliminating one or more documents opened by the user within the predetermined time period.

3. The method of claim 2, wherein determining the second set of documents associated with one or more collaborators based on events of one or more collaborators of the user and importance weights assigned to the one or more collaborators comprises: determining the one or more collaborators by identifying one or more other users of the cloud-based content management platform who provided at least one of a view event, an edit event, or a comment event to a document shared with the user in the cloud-based content management platform within the predetermined time period; determining an importance weight for each collaborator in a social network associated with the cloud-based content management platform based on the importance of the collaborator; as well as The second document set is determined, the second document set including one or more documents having events of the one or more collaborators and importance weights assigned to the one or more collaborators.

4. The method according to claim 2, wherein determining the third document set related to the documents in the document set of the user's work comprises: identifying key text for each document in a collection of documents of the user's work; and The third document set including one or more documents having the key text is determined. 5 . The method of claim 2 , wherein at least one of the second document set or the third document set includes one or more documents that have not been accessed by the user.

6. The method of claim 2, wherein generating the first unexpected document set comprises one of the following: Eliminating one or more documents for events to which the user does not have access and which have less than a predetermined number of one or more collaborators; or One or more documents that are set to be undiscoverable by searches performed via the cloud-based content management platform are eliminated.

7. The method of claim 1, wherein identifying the one or more target documents of the user comprises one of the following: determining a number of viewing events associated with the corresponding document by other users or the one or more collaborators; or The number of collaboration events is determined based on the number of interaction events between the user and the corresponding document, the number of interaction events between the one or more other users and the corresponding document, and the likelihood of the user performing an interaction event with one or more other users involved in the corresponding document, wherein an interaction event is an event generated in response to an event of another user.

8. The method of claim 1, wherein identifying the one or more target documents of the user comprises: For each document in the first document set, applying as input to a machine learning model an amount of topical overlap between the corresponding document and a document set of the user's work, a number of view events for the corresponding document, and a number of collaboration events associated with the corresponding document; and An output is obtained from the machine learning model, the output indicating a probability that the user is interested in the corresponding document.

9. The method of claim 1 , wherein the home screen GUI includes one or more suggestion cards, each suggestion card identifying each target document, and each suggestion card including: Title information, the title information including the document type and title of the corresponding document; a graphical representation of the corresponding document; an action button providing access to the corresponding document; as well as A reason text describing a reason for identifying the corresponding document.

10. A system comprising: A server of a cloud-based content management platform, the server comprising: memory devices; and a processing device operably coupled to the memory device, the processing device configured to: In response to a login request from a user to log into the cloud-based content management platform, one or more target documents that may be of interest to the user and have not been opened by the user or have been opened by the user for more than a predetermined period of time are selected from a plurality of documents associated with the user account of the user on the cloud-based content management platform, wherein selecting the one or more target documents comprises: identifying a first set of unexpected documents among the plurality of documents associated with the user account of the user, wherein the first set of unexpected documents associated with the user account of the user includes documents uploaded to the cloud-based content management platform by the user and documents uploaded to the cloud-based content management platform by other users, wherein each unexpected document in the first set is a document that has expired from cache or has been opened by the user for more than the predetermined time period; and identifying the one or more target documents from the first unexpected document set from the user based on: (i) an amount of topical overlap between unexpected documents from the first unexpected document set that have expired from cache or have been last opened by the user more than the predetermined time period and a document set of the user's work that was last opened within the predetermined time period, (ii) a number of view events for the unexpected documents, and (iii) a number of collaboration events associated with the unexpected documents; and In response to the login request of the user, a home screen graphical user interface (GUI) of the cloud-based content management platform is provided for presentation to the user, the home screen GUI identifying the one or more target documents.

11. The system according to claim 10, wherein: To identify the first unexpected document set, the processing device is further configured to: determining a second set of documents associated with the one or more collaborators based on events of the one or more collaborators of the user and importance weights assigned to the one or more collaborators; determining a third document set related to the documents in the document set of the user's work; as well as The first unexpected document set is generated by combining the second document set and the third document set and eliminating one or more documents opened by the user within the predetermined time period.

12. The system of claim 10, wherein in order to determine the second document set associated with the one or more collaborators based on events of the one or more collaborators of the user and importance weights assigned to the one or more collaborators, the processing device is further configured to: determining the one or more collaborators by identifying one or more other users of the cloud-based content management platform who provided at least one of a view event, an edit event, or a comment event to a document shared with the user in the cloud-based content management platform within the predetermined time period; as well as determining an importance weight for each collaborator in a social network associated with the cloud-based content management platform based on the importance of the collaborator; and The second document set is determined, the second document set including one or more documents having events of the one or more collaborators and importance weights assigned to the one or more collaborators.

13. The system according to claim 10, wherein in order to determine the third document set related to the documents in the document set of the user's work, the processing device is further configured to: identifying key text for each document in the document collection of the user's work; and The third document set including one or more documents having the key text is determined.

14. The system according to claim 10, wherein in order to identify the one or more target documents of the user, the processing device is further configured to do one of the following: determining a number of viewing events associated with the corresponding document by other users or the one or more collaborators; or The number of collaboration events is determined based on the number of interaction events between the user and the corresponding document, the number of interaction events between the one or more other users and the corresponding document, and the likelihood of the user performing an interaction event with one or more other users involved in the corresponding document, wherein an interaction event is an event generated in response to an event of another user.

15. The system of claim 10, wherein the home screen GUI includes one or more suggestion cards, each suggestion card identifying each target document, and each suggestion card including: Title information, the title information including the document type and title of the corresponding document; a graphical representation of the corresponding document; an action button providing access to the corresponding document; as well as A reason text describing a reason for identifying the corresponding document.

16. A non-transitory computer-readable medium storing instructions that, when executed, cause a processing device of a cloud-based content management platform to: In response to a login request from the user to log into the cloud-based content management platform, one or more target documents that may be of interest to the user and have not been opened by the user or have been opened by the user for more than a predetermined period of time are selected from a plurality of documents associated with the user account of the user on the cloud-based content management platform, wherein: Selecting the one or more target documents includes: identifying, by a server of the cloud-based content management platform, a first set of unexpected documents among the plurality of documents associated with the user account of the user, wherein the first set of unexpected documents associated with the user account of the user includes documents uploaded to the cloud-based content management platform by the user and documents uploaded to the cloud-based content management platform by other users, wherein each unexpected document in the first set is a document that has expired from cache or has been opened by the user for more than the predetermined time period; and The one or more target documents of the first unexpected document set from the user are identified by the server of the cloud-based content management platform based on: (i) an amount of topical overlap between unexpected documents from the first unexpected document set that have expired from cache or have been last opened by the user for more than the predetermined time period and a document set of the user's work that was last opened within the predetermined time period, (ii) a number of view events for the unexpected documents, and (iii) a number of collaboration events associated with the unexpected documents; and In response to the login request of the user, a home screen graphical user interface (GUI) of the cloud-based content management platform is provided by the server of the cloud-based content management platform for presentation to the user, the home screen GUI identifying the one or more target documents.

17. The computer-readable medium of claim 16, wherein: To identify the first set of unexpected documents, the instructions further cause the processing device to: determining a second set of documents associated with the one or more collaborators based on events of the one or more collaborators of the user and importance weights assigned to the one or more collaborators; determining a third document set related to the documents in the document set of the user's work; as well as The first unexpected document set is generated by combining the second document set and the third document set and eliminating one or more documents opened by the user within the predetermined time period.

18. The computer-readable medium of claim 16, wherein to determine the second set of documents associated with one or more collaborators of the user based on events of the one or more collaborators and importance weights assigned to the one or more collaborators, the instructions further cause the processing device to: determining the one or more collaborators by identifying one or more other users of the cloud-based content management platform who provided at least one of a view event, an edit event, or a comment event to a document shared with the user in the cloud-based content management platform within the predetermined time period; determining an importance weight for each collaborator in a social network associated with the cloud-based content management platform based on the importance of the collaborator; and The second document set is determined, the second document set including one or more documents having events of the one or more collaborators and importance weights assigned to the one or more collaborators.

19. The computer-readable medium of claim 16, wherein to determine the third document set related to the documents in the document set of the user's work, the instructions further cause the processing device to: identifying key text for each document in the document collection of the user's work; and The third document set including one or more documents having the key text is determined.

20. The computer-readable medium of claim 16, wherein the home screen GUI includes one or more suggestion cards, each suggestion card identifying each target document, and each suggestion card including: Title information, the title information including the document type and title of the corresponding document; a graphical representation of the corresponding document; an action button providing access to the corresponding document; as well as A reason text describing a reason for identifying the corresponding document.

21. The method according to claim 1, wherein: The home screen GUI presents a message indicating that the one or more target documents are suggested to the user based on serendipity.

Citation Information

Patent Citations

  • Display controller and operation method thereof

    US20180047370A1

  • Providing relevant text auto-completions

    CN101681198A

  • System and method for automating information abstraction process for documents

    CN106250385A

  • Aggregating and presenting recent activities for synchronized online content management systems

    US20160294916A1

  • Collaboration cards for communication related to a collaborated document

    US20170147547A1