• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    METHOD FOR SCREENING ELECTRONIC COMMUNICATIONS IN A COMPUTER SYSTEM, COMPUTING DEVICE, AND COMPUTER-READABLE STORAGE ENVIRONMENT. The present invention relates to screening electronic communications in a computing system environment that can mitigate problems relating to large volumes of incoming electronic communications communications. This may include analyzing user-specific electronic communication data and associated behavior to predict which communications the user is likely to deem important or unimportant. Client-side application capabilities are exposed based on the assessment of communication importance, to allow the user to arbitrarily process large volumes of incoming communication.
    公开号:BR112013012553B1
    申请号:R112013012553-5
    申请日:2011-11-20
    公开日:2021-06-08
    发明作者:Tore Sundelin；James Kleewein；James Edelen；Jorge Pereira；Alexander Wetmore；John Winn
    申请人:Microsoft Technology Licensing, Llc；
    IPC主号:
    专利说明:

    BACKGROUND
    [001] The growing use of electronic devices to manage personal and/or professional communications typically translates into an increase in incoming messages. In many cases, the sheer volume of incoming messages often precludes the end user's ability to effectively process it. Examples of problems and inefficiencies in the flow of such messages include an increasing potential for managing important messages and increased investment in the time required to filter incoming messages. SUMMARY
    [002] In one aspect, a method for screening electronic communications in a computer system environment includes: training a default model on a computing device to customize a specific container model for a container, the default model being formed from a plurality of weighting factors adjusted against a sample of users having common characteristics with the container, and the container-specific model being formed from the default model, which is modified using historical container behavior and feedback information; intercept an item addressed to the recipient on the computing device; extract a plurality of item resources associated with the item on the computing device; retrieve the specific container model; the container-specific model comprising the plurality of weighting factors associated with the plurality of extracted item resources; apply an importance ranking model to the plurality of extracted item features, including forming a combination of the plurality of weighted factors; generate a predicted item importance based on the combination of the plurality of weighted factors; and enable at least one application feature associated with the item for the container, based on the expected importance of the item.
    [003] In another aspect, a computing device includes: a processing unit; a system memory connected to the processing unit, the system memory including instructions, which, when executed by the processing unit, cause the processing unit to implement a training module to hierarchically train a user model for screening electronic communications in a computer system environment, the training module being configured to: generate a set of default inferences for a user based on the prototype user model, where a default inference comprises an item attribute; attribute value, attribute weight, and attribute confidence; acquire user-specific information to customize the default set of inferences for the user including: retrieving user-specific historical behavior and feedback information; and retrieving specific user behavior and feedback information in response to receiving an item; update the default set of inferences with user-specific information to form a custom set of inferences for application to an item screening template; and enable at least one application feature associated with the user to expose predicted item importance.
    [004] In yet another aspect, computer-readable media has computer-executable instructions that, when executed by a computing device, cause the computing device to perform the steps:
    [005] train a default model on the computing device to customize the specific container model for a container in which the default model is formed from a plurality of weighted factors, adjusted against a sample of users having common characteristics with the container, the common characteristics being selected from the group including: common calling; and common interest, and the container-specific model being formed from the default model that is modified using historical behavior and feedback information; intercept an item addressed to the recipient on the computing device; where the selected item from a group includes: email message, scheduled message; instant messaging, web-based messaging, and social collaboration messaging; extract a plurality of item resources associated with the item on the computing device; where the item features include a characteristic of the item selected from a group including: item issuer characteristic; conversation feature, and attachment features; retrieving the container-specific model wherein the container-specific model comprises the plurality of weighted factors associated with the plurality of extracted item resources; apply an importance ranking model to the plurality of extracted item features, including forming a combination of the plurality of weighted factors; generate a predicted item importance based on the combination of the plurality of weighted factors, with the predicted item importance designating the item as one of: important and unimportant; enable at least one app feature associated with the item for the container based on predicted item importance selected from a group including: an emphasizing feature to emphasize the key content of the item, a video feature to provide a quick preview of the item, and a notification feature to provide a temporary preview of the item; and periodically acquire the container behavior and feedback associated with the item for a predetermined period of time, to proceed with default model training to customize the specific container model.
    [006] This summary is provided to present a selection of concepts in simplified form, which will be further described below in the "Detailed Description" section. This summary is not intended to identify key or essential features of the claimed object, nor is it intended to be used in any way that limits the scope of the claimed object. DESCRIPTION OF DRAWINGS
    [007] Aspects of the present specification can be more fully understood from the following detailed description of various configurations, in connection with the accompanying drawings:
    [008] Figure 1 is a flowchart of an exemplary method for training user model data for screening electronic communications;
    [009] Figure 2 shows an example of a network computing environment;
    [0010] Figure 3 shows an exemplary server computing device of the Figure 2 environment;
    [0011] Figure 4 shows exemplary logic modules of a client device of the Figure 2 environment;
    [0012] Figure 5 shows an exemplary screening application environment;
    [0013] Figure 6 is a flowchart of an exemplary method for hierarchical training of user model data for screening electronic communications;
    [0014] Figure 7 shows a first view of an exemplary message sorting environment;
    [0015] Figure 8 shows a second view of the message sorting environment of Figure 7; and,
    [0016] Figure 9 shows a first view of another exemplary message sorting environment. DETAILED DESCRIPTION
    [0017] This specification is directed to systems and methods for screening electronic communications in a computer system environment. Screening techniques described in this alleviate issues relating to large volumes of incoming electronic communications, allow for an analysis of user-specific electronic communications data and associated behaviors to determine which communications a user is likely to regard as important or unimportant. Communication importance assessment is used to expose application capabilities that allow the end user to effectively process arbitrarily large volumes of incoming communications. Although not limited thereto, an appreciation of various aspects of the present specification will be gained by discussing the examples provided below.
    [0018] Referring now to Figure 1, an exemplary method 100 for training user model data in electronic communications screening is shown. In general, method 100 can be implemented by a server-side or a client-side process. Examples of server-side and client-side processes are described below with reference to Figures 2 to 9. Other configurations are possible. For example, method 100 can be implemented in a hybrid fashion, incorporating functionality from both server-side and client-side processes.
    [0019] Method 100 starts in a collection module 105. Collection method 105 is configured to retrieve electronic communication data from a recipient, such as an individual or a group of individuals, from a process that manages the communication data. Electronic communication data is often called an item. An exemplary item includes an email message, voice mail message (voicemail), schedule, SMS message, IM message, MMS message, web update, Facebook message, Twitter feed, RSS feed, electronic document, and others . Other configurations are possible.
    [0020] The operational flow proceeds to an analysis module 110. The analysis module 110 is configured to extract a plurality of item resources from the item, retrieved by the collection module 105. An item resource is generally any conceivable characteristic, which can be extracted or inferred directly based on understanding the content of the item.
    [0021] For example, an item resource may include a characteristic related to an issuer and/or recipient of the item, such as, for example, issuer/recipient identification (eg SMPT address), issuer/recipient relationship (by example, supervisor); issuer/recipient domain or company (eg Microsoft) issuer/recipient type (eg Automail), issuer/recipient location (eg emergency room); sender/recipient device (eg Smartphone), item shipping characteristics (eg CC), and others. Other exemplary item features include a feature related to recipient and/or contextual characteristics, such as, for example, present or future issuer/recipient status (eg, in meeting), location of present or future issuer/recipient (by example, Minneapolis), and others.
    [0022] Other exemplary item features include a feature related to an item type (eg email message), attachment presence (eg Yes), access control information (eg DRM), access control information. priority (eg High), temporal information (eg date/time received), and others. Other exemplary item features include a feature related to a conversation starter feature (eg started by me ), conversation contribution features (eg my contributions ), item hierarchical features (eg last in conversation ) and others. Other exemplary item features include a feature related to subject line prefix (eg RE), subject line passwords (eg READ) and others. Other exemplary item features include features relating to the item body or item attachment, such as, for example, text keywords (eg important), hyperlinked content (eg Yes - contains hyperlink), and others.
    [0023] Still other item features are possible.
    [0024] The operational flow then proceeds to an acquisition module 115. The acquisition module 115 is configured to retrieve specific template data for each intended container of the item, as retrieved by the collection module 105. In the exemplary discussion that follows follows, the intended container includes a single individual, and the container-specific model data is retrieved by the acquisition module 115 from a data storage device. An exemplary data storage device is described below in connection with Figure 2.
    [0025] In exemplary configurations, the container-specific model data includes a plurality of item resources (e.g., corresponding to item resources extracted by the analysis module 110), each of which is assigned an embedding weight. the indication as to whether the recipient tends to attach importance or unimportance to the respective item resource. For example, if the container tends to read email messages sent from a supervisor, and ignore email messages sent from an automated system, an item resource within the template data for the container associated with the supervisor could include a weight factor greater than an item resource associated with the automated service. In general, a weight or weighting factor can include any form of quantitative measure, such as a numerical value, a threshold, and so on. For example, an item resource associated with the supervisor, as discussed above, could include a weight of "7", while the item resource associated with the automated service could include a weight of "3".
    [0026] The operational flow then proceeds to an implementation module 120. The implementation module 120 is configured to apply a model criterion from a classification model to a specific container model data as retrieved by the acquisition module 115. How described in detail below with respect to Figure 6, container-specific model data can be formed via a hierarchical training process, using prototype model data that has been calculated by analyzing training data from a number of related users, for more efficiently and precisely train a model for a user (eg the container). Other configurations are possible.
    [0027] The implementation module 120 is further configured to generate one or more predictions, based on the type of classification model. An exemplary model criterion includes the assignment of item resources from the container-specific model data that are relevant to the classification model, and, additionally, an algorithm that designates the use of weights associated with those item resources evaluated as relevant.
    [0028] In exemplary configurations, the classification model corresponds to an "importance" model, where implementation model 120 correlates relevant item resources from container-specific model data to associated weights; and uses a combination of these weights to generate a predicted item importance. Predicted item importance generally includes prediction with respect to whether the item, as retrieved by the collection module 105, is important or unimportant to the intended recipient. Other configurations are possible. For example, in some configurations, the classification model corresponds to the "urgency" model, where implementation module 120 correlates relevant item resources from container-specific model data to associated weights, and uses a combination of those weights to generate an anticipated item urgency, indicating the items the intended recipient should consider or attend to as soon as possible. Still other configurations are possible.
    [0029] An exemplary application of an importance model includes calculating the overall importance weight of a new email message, and then determining whether or not the email message is important to the recipient, based on the calculated importance weight. For example, on a scale of 1 to 10, a calculated importance weight of "4" indicates that the email message is "moderately important", a calculated importance weight of 7 to 8 indicates that the email message is " extremely important", and an importance weight of less than 6 indicates that the email message is "unimportant". Other configurations are possible. For example, in some configurations, an overall importance weight for a new email message is calculated as a probability ranging from 0 to 1, to indicate the relative importance of the email message. For example, thresholds ranging from 0 to 0.2 indicate that the relevant importance of the email message is unimportant or cold, thresholds ranging from 0.2 to 0.8 indicate that the relative importance of the email message is normal , and thresholds ranging from 0.8 to 1 indicate that the relative importance of the email message is important or warm. Still other configurations are possible.
    [0030] The operational flow then proceeds to the storage module 125. The storage module 125 is generally configured to store model-specific container data, as retrieved by the acquisition module 115, and the one or more predictions generated by the storage module. implementation 120.
    The operational flow then branches between a first training lead 130 and a second training lead 135. The first training lead 130 includes a first monitoring module 140 and a first extraction module 145. The second training lead 135 includes a second monitoring module 150 and a second extraction module 155. In general, the operational flow within the first training lead 130 is independent of the second training lead 135.
    [0032] Referring now to the first training lead 130, the first monitoring module 140 is configured to monitor and acquire container behavior with respect to the item, as retrieved by the collection module 105. The exemplary container behavior includes any form of action relating to the directly observable item. Such observable actions can be either a singular or a composite action. In the example of an email message, the recipient behavior can be associated with singular actions, such as opening, deleting, forwarding an email message. Composite actions can include actions such as quickly scanning the email message and then deleting it, denying access to an email message automatically archived to a file via a transport rule, and others.
    [0033] The first monitoring module 140 is configured to monitor and acquire container behavior with respect to the item, as retrieved by collection module 105, for a pre-determined period of time dT. An exemplary time period includes a fraction of an hour, an hour, a day, a week, etc. Following the expiration of the predetermined time period dT, the first monitoring module 140 forwards the acquired container behavior to the first extraction module 145. In other configurations, the first monitoring module 140 is further configured to monitor and acquire the behavior of a container with respect to the item, as retrieved by the collection module 105 on the basis of a container action, assigning a relative importance or following the course of a predetermined period of time, whichever occurs first. Examples of container actions that designate relative importance include "responded" to designate importance, "quickly debugged and promptly deleted" to designate unimportance, and others. Acquiring a container's behavior based on the combination of container action and the passage of a pre-determined period of time allows for a quick and efficient update of the classification model, as described in further details below.
    [0034] The first extraction module 145 is configured to mine acquired container behavior and generate behavior verification data. In general, the behavior verification data contains information regarding whether the predicted item importance generated by the implementation module 120 is consistent with whether the recipient actually judges the item to be important or unimportant. The first extraction module 145 subsequently forwards the behavior verification data to an update module 160. The update module 160 is configured to adjust weights associated with the plurality of item resources of the container-specific model data. For example, in the example of an email message, if the behavior verification data contains information that strongly suggests that the recipient considers email messages from a supervisor important, an item resource associated with the supervisor, such as discussed above, could be adjusted or readjusted from a "7" weight to a "9" weight. Other configurations are possible.
    [0035] In exemplary configurations, the operational flow returns to the first monitoring module 140 from the first extraction module 145, following a predetermined time delay dT. The closed-loop process flow within the first training lead 130 serves to continuously and precisely adjust container-specific model data based on container actions.
    [0036] Referring now to the second training lead 135, the second monitoring module 150 is configured to monitor and acquire a container return relative to item importance, as retrieved by collection module 105. The exemplary container return includes any form of explicit return from a recipient relative to the item's importance. In the email message example, the explicit return may include a recipient correction of the predicted item importance generated by the implementation module 120, such as marking the email message as unimportant when the implementation module 120 has incorrectly named as important in the email message. Other configurations are possible.
    [0037] For example, another explicit return includes enabling or disabling certain processing rules or calibration of processing rules relating to importance, such as disabling the use of an issuer's company as an indicator of the item's importance. Another explicit return includes setting thresholds with respect to levels of importance, such as setting items as important only when the relative importance is greater than a threshold weight. Other explicit feedback includes customizing existing processing rules or defining new processing rules, such as naming a spouse email message including the string "911" as urgently important. Still other configurations are possible.
    [0038] The second monitoring module 150 is configured to monitor and acquire container return with respect to the item, as retrieved by the collection module 105, for a predetermined period of time dT (i.e., one hour, one day, one second, etc.). Following the expiration of the predetermined time dT, the second monitoring module 150 forwards the acquired container return to the second extraction module 155. Other configurations are possible.
    [0039] The second extraction module 155 is configured to mine the acquired container return, and generate return verification data. In some configurations, the checkback data contains explicit designation with respect to whether the predicted item importance generated by the implementation module 120 is consistent with respect to whether the recipient actually deems the item important or unimportant. The second extraction module 155 subsequently forwards the verification data back to the update module 160. In the exemplary case, the update module 160 is configured to adjust the weights associated with the plurality of item resources of the container-specific model data with base on container return. For example, in the example of an email message, if the bounce verification data contains the designation that the recipient strongly considers to be unimportant email messages coming from an automated service, the item resource associated with the service Automated, as discussed above, could be adjusted from a "5" weight to a "1" weight. Other configurations are possible.
    [0040] The operational flow returns to the second monitoring module 150 from the second extraction module 155, following a predetermined time delay dT. The closed loop process flow within the second training tap 135 serves to continuously fine-tune the vessel-specific model data based on vessel feedback.
    [0041] In some configurations, the container-specific model data is updated differently based on information received via first instrument lead 130 and second training lead 135 respectively. For example, the confidence associated with the information received by the second training lead 135 can be assigned a greater confidence than the information received by the first training lead 130. In this way, the information received by the second training lead (i.e., explicit feedback) will have a greater impact on the training of the container-specific model data than the information received by the first training lead 130 (ie, implicit feedback). For example, in some configurations, the information received by the second training lead 135 completely overrides the information received by the first training lead 130. Other configurations are possible.
    [0042] In addition, the information received by the first training lead 130 can be assigned varying strengths with respect to confidence in relation to importance, to determine which information exerts the greatest impact on the training of the container-specific model data. For example, an observed container action, such as "respond", can be assigned a force greater than "read from length", which may be assigned a force greater than "ignored", which may be assigned a force greater than "quickly read", etc. Other configurations are possible.
    [0043] As will be described in more detail below with respect to Figures 2 through 9, the exemplar method 100 allows for a wide variety of client-side application features so that the end user can effectively screen large volumes incoming communication. A client-side application feature includes an emphasizing feature to emphasize key content in an item. Such emphasizing feature has little impact, as it clearly marks certain items or inserts content into a communication, to help the user quickly triage the communication, but without substantially changing the functionality of the client-side application.
    [0044] Another example of a client-side application feature includes a quick view feature that allows the user to quickly view only the most important communications. Another client-side application feature includes an auto-prioritization feature that provides a view sorted according to the most important communications. Another client-side application feature includes an obsolescence feature, which automatically archives, marks read, or erases unwanted communications after a certain period of time. Another example of a client-side application feature includes a notification feature, which is configured to selectively provide new communications and/or content notifications based on communications deemed important. In some configurations, the notification feature is user context sensitive. Another exemplary client-side application feature includes a synopsis feature, which provides a synopsis of the communication's content to help the user quickly decide what actions to take with respect to the communication. Another example of a client-side application feature includes a command panel feature, which provides a consolidated view of important communications from various data sources, such as email data source, document data source, Web-based data source, social network data source.
    [0045] Still other client-side application features are possible.
    [0046] Referring now to Figure 2, an exemplary network computing environment 200 is shown, in which aspects of the present specification can be implemented. The network computing environment 200 includes a client device 205, server device 210, storage device 215, and network 220. Other configurations are possible. For example, the networked computing environment 200 can generally include more or less devices, networks, and other components, as desired.
    [0047] The client device 205 and server device 210 are general purpose computing devices as described below in connection with Figure 3. In exemplary configurations, the server device 210 is an enterprise server that implements business processes. Exemplary business processes include messaging processes, collaboration processes, data management processes, and others. Microsoft Corp's "Exchange Server" is an example of an enterprise server that implements messaging and collaborative business processes that support electronic mailing, scheduling, and task contact features, to support mobile web-based access to information, to support storage of data. SHAREPOINT®, also from Microsoft Corp, is an example of a business server that implements business processes to support collaboration, file sharing, and Web editing. Other business servers, which implement business processes, are also possible.
    [0048] In some configurations, server device 210 includes a plurality of interconnected server devices operating together in a "Park" configuration to implement business processes. Still other configurations are possible.
    [0049] Storage device 215 is a data storage device, such as a relational database, or any other type of persistent data storage device. Storage device 215 stores data that server device 210 can fetch, modify, and manage. Examples of data storage devices include storage and addressing services such as ACTIVE DIRECTORY® from Microsoft Corp. Other configurations of storage device 215 are also possible.
    [0050] Network 220 is a bi-directional data communication path to transfer data between one or more devices. In the example, network 220 establishes a communication path for transferring data between client device 205 and server device 210. In general, network 220 can be any network of a number of wired or wireless WAN, LAN, Web, or wireless networks wired or other packet-based, wired or wireless communication networks so that data can be transferred within the elements of the networked computing environment 200. Other network configurations are also possible.
    [0051] Referring now to Figure 3, the server device 210 of Figure 2 is shown in greater detail. As mentioned, server device 210 is a general-purpose computing device. General purpose computing devices include desktop computer, laptop computer, Personal Data Assistant (PDA), Smartphones, Servers, Netbook, Notebook, Tablet, Cell Phone, Television, Video Game Console, and others.
    [0052] Server device 210 includes at least one processing unit 305 and system memory 310. System memory 310 can store and operate system 315 to control the operation of server device 210 or any other computing device. A 315 operating system is Microsoft Corp's WINDOWS® operating system or a Server such as Exchange Server, SHAREPOINT® collaborative server and others.
    [0053] System memory 310 also includes one or more software applications 320 and may include program data. Software applications 320 may include different types of single or multi-functional programs, such as electronic mail program, scheduling program, and Internet search program, spreadsheet program, information tracking and reporting program, information processing program. text, and many others. An example of a multifunctional program is Microsoft Corp's OFFICE suite.
    [0054] System memory 310 may include computer readable storage media, for example, a magnetic disk, optical disk, or tape. Such additional storage is illustrated in Figure 3 by removable storage 325 and non-removable storage 330. Computer readable storage media may include volatile and non-volatile, removable and non-removable physical media implemented by any method or technology of storing information, such as computer readable instructions, data structure, program modules, or other data. Computer readable storage media may also include without limitation RAM, ROM, EEPROM, Flash Memory, or other memory technologies, CD-ROM, digital versatile disks (DVD), or other magnetic storage devices, or any other media that they can be used to store desired information and be accessed by the server device 210. Any such computer storage media may be part of, or may be external to, a server device 210.
    [0055] Communication media is different from computer readable storage media. Communication media typically can be configured using computer readable instructions, data structure, program modules, or other data in a modulated data signal, such as a carrier wave or other transport mechanism, and includes any media. of information supply. The term "modulated data signal" refers to a signal having one or more of the characteristics, which are adjusted or changed so as to encode information in the signal. By way of example, communication media includes wired media or direct wired connection and wireless media such as acoustic media, RF media, infrared media, etc.
    [0056] Server device 210 may also have any number and type of input devices 335 and output devices 340. An exemplary input device 335 includes keyboard, mouse, pen, voice input device, touch input device, and others. The exemplary output device 340 includes video monitor, speakers, printers and others. Server device 210 may also include a communication link 345 configured to allow communication with other computing devices via a network (eg, network 220 of Figure 2) in a distributed operating system environment.
    [0057] In exemplary configurations, the client device 205 of Figure 2 is configured similarly to the server device 210 described above. Referring further to Figure 4, client device 205 of Figure 2 is also configured to include one or more different types of client interfaces to server device 210. In the example shown, client device 205 includes a local device 405 , a web access client, a mobile access client 415, and a voice access client 420. Other types of client interfaces to the server device 210 are also possible.
    [0058] The local client 405 is configured as a collaborative client and dedicated messaging client, which interfaces with the server device 210 and is part of a suite of applications on the client device 205. In one configuration, the local client 405 includes a OUTLOOK® messaging client which is an email application that is part of the Microsoft Office suite. A user can compose, interact, send, and receive emails with OUTLOOK®. Other 405 local client configurations are also possible.
    [0059] Web access client 410 is configured to access server device 210 remotely, using a network connection, such as the Internet. In one configuration, the Web Access Client 410 is the Outlook Web Access Web mail service. In the exemplary configuration, the client device 205 uses a web browser to connect to Exchange Server via Outlook Web Access. This provides an interface similar to the interface in the OUTLOOK® messaging client in which the user can compose, interact, send and receive emails. Other 410 web access client configurations are also possible. For example, Web Access Client 410 can be configured to connect to SHAREPOINT® Collaborative Server to access the corresponding Collaboration, File Sharing, and Web Publishing Service. Still other Web Access Client 410 configurations are possible.
    [0060] Mobile access client 415 is another type of client interface with device server 210. In one configuration, mobile access client 415 includes mobile access with ACTIVESYNC® Windows Mobile Device Center for VISTA synchronization software or WINDOWS 7, all from Microsoft Corp. A user can synchronize messages between a mobile device and Exchange Server using a mobile access client like Mobile Access with ACTIVESYNC® synchronization software. Exemplary mobile devices include cell phone, Personal Digital Assistant, and others. Other 415 mobile access client configurations are also possible.
    [0061] Voice access client 420 is yet another type of client interface with server device 210. In some configurations, voice access client 420 includes Exchange Unified Messaging, supported on Exchange Server. With Exchange Unified Messaging, users have inbox for email and voice mail (voicemail). Voice messages are delivered directly to the OUTLOOK® message client inbox. Messages containing voice messages also include attachments. Other 420 voice access client configurations are also possible.
    [0062] Referring to Figure 5, an operating system environment 500 configured to implement systems and methods for screening electronic communications in an operating system environment is shown. The operating environment 500 can be implemented by a server-side process running on the server computing device or by a client-side process running on a client computing device, as described above in connection with Figures 1 to 4 Other configurations are also possible. For example, operating environment 500 can be implemented in a hybrid fashion, incorporating functionality from both server-side and client-side processes. Such flexibility in implementing exemplary systems and methods for screening electronic communications is advantageous in many respects, such as, for example, allowing for optimal resource allocation, load balancing, and others.
    [0063] The exemplary operating environment 50 includes a data collector 505, a data analyzer 510, a data store 515, and a fetch analyzer 520.
    [0064] The 505 data collector is configured to collect and aggregate raw item data from a variety of electronic communication and related sources, such as email data, voice data, social network data, electronic document data , and others. As electronic communications and related sources typically group and transmit data in different formats, data collector 505 may include multiple server data modules that support these differences, such as communication server data collector 525, data collector 525. web server 530, and application server data collector 535. Other types of data collector modules are also possible.
    [0065] The data analyzer 510 includes an item analyzer module 540, a model 545 application module and a data training module 550. The item analyzer module 540 is configured to extract a plurality of item resources from the respective item data, as retrieved by data collector 505. As discussed above, in the context of exemplary method 100, an item resource is generally any characteristic of communication data, which can be directly extracted or inferred based on an understanding of the contents of the respective communication data.
    [0066] The model 545 application module is configured to retrieve container-specific model data from the intended container that corresponds to the respective item data, as retrieved by the 505 data collector. The model 545 application module is further configured to apply a model criterion from a classification model to the container-specific model data and generate one or more item-specific predictions based on the classification model type. In one configuration, the ranking model is an importance-based model. Other configurations are also possible.
    [0067] Data training module 550 is configured to monitor and acquire container behavior and explicit container return associated with an intended container, which corresponds to the respective item data, as retrieved by data collector 505. 550 data training is additionally configured to adjust the container-specific model data, as retrieved by the model 545 application module, based on acquired container behavior and explicit container return.
    [0068] As mentioned above, the operating environment 500 also includes a search analyzer 520. In general, the search analyzer 520 is configured to process client-side application resource requests so that the end user can effectively perform the screening in large volumes of incoming communications. In the exemplary configuration, search analyzer 520 includes a first resource portal 555, a second resource portal 560, and a third resource portal 565.
    [0069] The first resource portal 555 is configured to support resource requests that correspond to an emphasizing resource to expose key content in a client-side application, as described in further detail below in connection with Figures 7 and 8 The second resource portal 560 is configured to support resource requests that correspond to a quick view feature to provide a quick view of items in a more important client-side application, also described below in connection with Figures 7 and 8. The third resource portal 565 is configured to support resource requests that correspond to a notification resource or provide selective notification within a client-side application based on items deemed important, described in additional detail below, in connection with Figure 9. Other 520 search analyzer configurations are also possible.
    [0070] In exemplary configurations, data collected in data collector 505 and/or processed by data analyzer 510 can be stored in data store 515. In addition, data store 515 supports and stores the searches and results processed by the 520 search analyzer.
    [0071] Referring now to Figure 6, an exemplary method 600 for providing a hierarchical training of user model data for screening electronic communications is shown. In general, method 600 can be implemented by a server-side or a client-side process. Examples of server-side and client-side processes are described above, in connection with Figures 1 to 5. Other configurations are also possible. For example, method 600 can be implemented in a hybrid fashion, incorporating the functionality of both server-side and client-side processes.
    [0072] Method 600 is configured to provide an optimal understanding of customer-specific behavior and preferences, which are called user model data. Exemplary inferences, according to the present specification, of a particular item attribute based on observed user behavior and explicit user feedback. In a configuration, an inference comprises an item attribute, attribute value, attribute weight, and attribute confidence. An exemplary set of user inferences that can be derived from user-specific behavior and feedback communication data includes:

    [0073] An item attribute of an inference is a characteristic of a particular communication. Examples of item attributes include issuer relationship, contain follow-up, and item topic. Other configurations are possible. For example, other attributes include item issuer, item topic, item ship time, item type, and others. The importance of a particular attribute value (eg, issuer = "manager" ratio) is assessed by looking at user behavior with respect to a particular attribute value. Exemplary user behavior may include looking at whether a user tends to exhibit behavior that indicates importance (for example, investing a significant amount of time with respect to the open item) for items coming from a manager. In the example shown, the attribute weight is represented by a numerical value on a scale. Other configurations are also possible.
    [0074] The confidence rating of an inference corresponds to the confidence associated with a particular importance rating of a given item attribute value. In the example shown, the confidence index of the inference "issuer ratio" = "manager" is relatively high (ie, 78%). The confidence index of the topic inference item = prospect is relatively low (ie, 23%). In some configurations, a high inference confidence index can be achieved when many attribute value cases are observed, associated with consistent behavior. A low inference confidence index can be achieved when few attribute value cases are observed, non-recent attribute value cases, and/or inconsistent user behavior. Other configurations are also possible.
    [0075] In some configurations, certain inferences are code-pending or composed of multiple related item attributes. An example of codependent inference includes a scenario in which a user receives many email messages from a colleague named "Alex". Some of these exemplary email messages are sent to a referral list (DL) where the user is listed, and others are sent directly. In one scenario, when "Alex" sends an email message to the user via the distribution list, the user tends to treat those items as unimportant. There are two related code-pending inferences that represent the exemplary scenario. A first inference includes attributes "issuer" = "Alex" and "recipient" = "DL", and may have a relatively low attribute weight (eg 4) and a high confidence index (eg 80%). A second inference includes "issuer" = "Alex" and "container" = "container", and may have a relatively high attribute weight (eg 8) and high confidence index (eg 80%). In general, any arbitrary codependent inference can comprise any number of arbitrary composite attributes.
    [0076] In exemplary configurations, compiling and calculating a set of weights for a particular user is called training. Exemplary method 69 is configured to train user model data in multiple stages. Specifically, operation 605 corresponds to a first-stage initialization operation, which generates a set of generalized default weights for a new user, based on a prototype user model. The default weight set represents user model data. An exemplary prototype user module includes an important model that has been developed, prototyped, and tested against a large population of users in the sample having common characteristics, such as common vocation, common interest, etc.
    [0077] Following the first stage bootstrapping in operation 605, user specific information is obtained to customize and adjust the default weight set to customize the default user template for a specific user. For example, the operational flow advances to operation 610 which corresponds to a second stage operation, which evaluates historically recorded available behavioral data, and communication data. Historically recorded behavioral data includes email message composition behaviors, such as sending, replying, and forwarding messages. Other historically recorded behavioral data and reporting items are possible and may be system implementation specific.
    [0078] Following the second stage "crawling" in operation 605, the operational flow goes to operation 615, which corresponds to updating the set of generalized default weights from the default user model data to form a custom set of weights. The custom set of weights matches the user-specific model data.
    [0079] Following the formation of user-specific model data in operation 615, the operational flow advances to a third-stage online operation 620, which corresponds to real-time monitoring and acquisition of specific behavior and feedback with respect to items similar to the functionality of the exemplary first training lead 130, described above in connection with Figure 1. Operation 620 is implemented to update and adjust the custom set of weights formed in operation 615.
    [0080] In exemplary configurations, the operational flow returns to operation 615, following a predetermined bit time delay. The closed-loop process flow between operations 615 and 620 is implemented to continuously fine-tune the custom set of user-specific model data weights. Such a closed loop process flow is advantageous in several respects. For example, certain weights can change or become obsolete over time, such as when the user changes jobs or changes supervisors. In the exemplary configuration, information obtained from operation 620 and corresponding updates from operation 615, capture the respective changes, and adapt the custom set of weights over time.
    [0081] In the exemplary configurations, the operational flow goes to evaluation operation 625, following interaction between operations 615 and 620. The evaluation operation 625 corresponds to the determination of whether the custom set of user-specific model data weights is sufficient to expose functionality based on associated ratings, such as triage features relating to marking new items as important.
    [0082] When evaluation operation 625 determines that the custom set of user-specific model data weights is insufficient to expose functionality based on the associated ratings, the operational flow returns to operation 620 to provide an adjustment and additional tuning of the custom set of weights.
    [0083] When evaluation operation 625 determines that the custom set of user-specific model data weights is sufficient to expose a functionality based on the associated ratings, the operational flow goes to operation 630, which corresponds to training completion start of the custom set of weights. In the exemplary configuration, triage features related to the associated classification are enabled in operation 630 and accessible through client-side application resource requests so that the end user can effectively screen arbitrarily large volumes of incoming communications (for example, first resource portal 555, second resource portal 560, and third resource portal 565). In general, initial training completion of the custom set of weights in the 630 operation can be achieved without any active or direct user action.
    [0084] Referring now to Figure 7, a first exemplary message environment 700 is shown, in accordance with the present specification. In general, the 700 messaging environment is an email application associated with a communication application such as OUTLOOK®. Other configurations are also possible.
    [0085] In exemplary configurations, the 700 message environment includes a folder pane 705, list panel 710, and a glass panel 715. Folder pane 705 includes a list of folders 720a to 720c used to store data, such as like email messages. In the example shown, folder 720c is selected to be represented on a list panel 710, such as email message list 725a to 725e.
    [0086] In the example shown, email message 725a is emphasized by a first importance mark 730, and email message 725b is emphasized by a second importance mark 735. In general, the first importance mark 730 designates email message 725a that is important because it comes from "Sheila Wu". Additionally, the email message 725a can be represented on the video panel 715, in a quick first view 740, because it is important. In the exemplary configuration, the first quick view 740 is configured to represent the 745 content of email message 725a and image 750 of "Sheila Wu". The geometry and tone of the first importance mark 730 is configurable and designates the presence of key content in the email message 725a, such as the term "Review" in the subject line. Other configurations are also possible.
    [0087] The second importance mark 735 designates email message 725b as important because it comes from "José Santana". The email message 725b can be represented on the pane display 715in a second quick view 755 because it is important. In the exemplar configuration, the second quick view 755 is configured to bring up the 760 content of email message 725b. The geometry and tone of the first importance mark 735 is configurable and can designate the presence of key content of the email message 725b, such as by having the term "Shipment" in the body text.
    [0088] In general, the first importance tag 730 and the second importance tag 735 allow the user to quickly identify email messages 725a and 725b as important. The geometry and tone of the first importance mark 730 and second importance mark 735b can be selected as desired and designate certain characteristics of the respective email messages 725a and 725b, and additionally influence the placement and relevance of the first quick view 740 and second quick view 755 on video panel 715, as desired. Other configurations are also possible.
    [0089] Referring now to Figure 8, the message environment 700 of Figure 7 is shown including a user module 800. In the exemplary configuration a cursor 805 is used to select the first importance tag 730 to expose to the user module 800.
    [0090] In general, user module 800 is configured to provide a high level of transparency to a user, to enable feedback and customization. For example, user module 800 can expose a set of user inferences 805a to 805c in a context sensitive and intuitive way. Exemplary user inferences provide an understanding of how the rating of a certain item (ie, email message 725a) is determined to be important or unimportant. The 800 user module is additionally configured to expose a manual adjustment knob 810, which allows the user to change the rating of an item from important to unimportant, if desired. In exemplary configurations, such active return updates the classification of that item as well as associated user template data, and additionally may account for the active return when sorting new items or reclassifying existing items.
    [0091] User module 800 is additionally configured to expose an inference feedback button 815 that allows the user to provide feedback by designating an inference as incorrect, or that a generally correct inference has been improperly applied to a particular email message. . Such active return serves to update that item's item rating as well as associated user template data. Such active return may additionally be received when classifying new items or reclassifying existing items. Inference return button 815 is additionally configured to allow the user to define new inferences or define special meta-inferences or a codependent inference calculated based on multiple attributes and values. Other configurations are also possible. User module 800 is additionally configured to expose a customize button 820 for user customization. An exemplary user customization includes threshold customization, such as defining the importance and/or confidence index or minimum importance that an item should be marked as important within the 700 messaging environment. Other exemplary user customization includes stratification, such as setting how many levels of importance should be assigned to a 700 message environment (eg, low, medium, high). Another exemplary customization includes defining a visual indicator to denote the relative importance in the 700 message environment (eg via icons, forecasts). Another exemplary user customization includes toolbar definition, to allow the user to define buttons or commands to expose application/features. Such customization can be device or application specific. Other exemplary user customization includes callback definition granularity, to allow the user to decide which active callback level controls to expose via the 700 message environment. Still other configurations are possible.
    [0092] Referring now to Figure 9, a second exemplary message operating environment 900 is shown, in accordance with the present specification. In general, operating environment 900 is a notification application, such as a "Smart Toast" window message (pop-up) application from Microsoft Corp. Other configurations are possible.
    [0093] In exemplary configurations, the message environment 900 is exposed to a user, with the receipt of a new email message evaluated as important. For example, similarly to the respective email message 725a described above, in connection with Figures 7 and 8, upon receipt of an email message from Sheila Wu, the message environment 900 can be represented by a pre- certain period of time, including the first importance mark 730 and image 750 of Sheila Wu. Message environment 900 also includes 905 metadata such as "New messages from Sheila Wu" and contextual 910 metadata such as message "will not be available until 14:00". In exemplary configurations, the 900 message environment notifies a user only those messages (in a 725a email message configuration) rated important, and quickly shows why those messages were rated important. Other configurations of message environment 900 will also be possible.
    [0094] The exemplary configurations described above can be implemented as logical operations on a computing device in a networked computing system environment. Logical operations can be implemented as: a computer-implemented sequence of instructions, steps or program modules executed in a computational device; and (ii) interconnected logic or hardware modules running in a computing device.
    [0095] For example, logical operations can be implemented as algorithms, firmware, analog/digital circuitry, or any combination thereof, without departing from the scope of this specification. The software, firmware, or similar sequence of program instructions can be encoded and stored on computer-readable storage media, and can also be encoded into a carrier wave signal for transmission between computing devices.
    [0096] Although the subject matter of this specification has been described in a specific language with respect to structural resources and/or methodological acts, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific resources or acts described above . Instead, specific remedies and acts were described as an exemplary way of implementing the claims.
    权利要求:
    Claims (20)
    [0001]
    1. Method (100) for screening electronic communications in a computing system environment characterized by the fact that it comprises the steps of: training a default model on a computing device to customize a specific container model for a container, where the default model is formed from a plurality of weighted factors adjusted against a sample of users having common characteristics with the container, and the container-specific model being formed from the default model which is modified using the historical behavior and feedback information of the container; intercepting an item addressed to the container in the computing device; extracting (110) a plurality of item resources associated with the item in the computing device; retrieving (115) the container-specific model, wherein the container-specific model comprises the plurality of weighted factors associated with the plurality of extracted item features; apply (120) a ranking model of importance to the plurality of extracted item features, including forming a combination of the plurality of weighted factors by calculating an importance weight as a probability range of threshold values; generating a predicted item importance based on the combination of the plurality of weighted factors; and enable at least one application feature associated with the item for the container based on predicted item importance.
    [0002]
    2. Method according to claim 1, characterized in that the common characteristics comprise one or more of a common vocation and common interest.
    [0003]
    3. Method according to claim 1, characterized in that it further comprises adjusting the plurality of weighted factors based on the historical behavior and return information of the recipient.
    [0004]
    4. Method according to claim 1, characterized in that it further comprises continuing training the default model to customize the specific container model by acquiring the behavior of the container associated with the item.
    [0005]
    5. Method according to claim 1, characterized in that it further comprises continuing training of the default model to customize the specific model of container by acquiring the return of the container associated with the item.
    [0006]
    6. Method according to claim 1, characterized in that it further comprises continuing training of the default model to customize the specific container model by acquiring the container customization, the container customization comprising one or more of: inference correction, processing rule definition, threshold definition, and importance granularity.
    [0007]
    7. Method according to claim 1, characterized in that it further comprises continuing training of the default model to customize the specific container model by periodically acquiring the container behavior associated with the item.
    [0008]
    8. Method according to claim 1, characterized in that it further comprises the importance of the item provided for designating the relative importance of the item.
    [0009]
    9. Method according to claim 8, characterized in that it further comprises periodically acquiring the container behavior associated with the item for a predetermined period of time to assess the correctness of the predicted item importance.
    [0010]
    10. Method according to claim 9, characterized by the fact that further comprising adjusting at least one of: plurality of weighted factors and the predicted item importance based on the acquired container behavior.
    [0011]
    11. Method according to claim 8, characterized in that it further comprises periodically acquiring the return of the container associated with the item for a predetermined period of time to assess the regularity of the importance of the anticipated item.
    [0012]
    12. Method according to claim 11, characterized in that it further comprises adjusting at least one of the following: the plurality of weighted factors; and the importance of the predicted item based on the recipient's return.
    [0013]
    13. Method according to claim 1, characterized in that the item includes a communication selected from a group including: e-mail message, voice message (voicemail), scheduled message, instant message; Web-based messaging, and social collaboration messaging.
    [0014]
    14. Method according to claim 1, characterized in that the extracted item features include at least one of a directly observed item feature and inferred item feature.
    [0015]
    15. Method according to claim 1, characterized in that it further comprises enabling the application features selected from the group that includes: an emphasis feature to highlight key content of the item; a display feature to provide a quick view of the item; a notification feature to provide temporary viewing of the item and including information related to the importance derived from the item; a self-prioritization feature to provide a ranked importance view of the item and other items; an age-out feature to provide an action for the item after a period of time; a synopsis feature to provide synopsis of the item's content; and a dashboard feature to provide a consolidated view of important communications between different data sources.
    [0016]
    16. A computing device comprising: a processing unit (305); a system memory (310) connected to the processing unit, the system memory including a method for implementing a training module configured for hierarchical training of a user model for screening electronic communications in a computer system environment, the training module characterized by the fact that it is configured to: generate a set of default inferences for a user based on a prototype user model, where the default inference comprises an item attribute, attribute value, attribute weight, and attribute confidence; acquire user-specific information to customize the default set of inferences for the user, including: user-specific historical behavior retrieval and feedback information, and retrieval behavior and user-specific feedback information in response to receiving an item; using the default set of inferences with user-specific information to form a custom set of inferences for application to an item screening model; and, enable at least one user-associated application feature to expose a predicted item importance, the predicted item importance being generated from an importance ranking model used to calculate an importance weight as a probability range of values limits based on a combination of a plurality of weighted factors.
    [0017]
    17. Device according to claim 16, characterized in that an item comprises an electronic communication, and in which the item attribute comprises a characteristic of a particular communication element, the attribute value comprises a specific instance of the attribute of item, the attribute weight comprises a scale value denoting importance of the attribute value, and the attribute confidence comprises a value designating confidence associated with the attribute weight.
    [0018]
    18. Device according to claim 16, characterized in that the prototype model comprises a plurality of weighted factors adjusted against a sample of users having common characteristics with the user, the common characteristics comprising one or more of a common vocation and of common interest.
    [0019]
    19. Device according to claim 16, characterized in that the retrieval of specific user behavior and feedback information in response to receipt of an item comprises the acquisition of periodic data to continuously adjust the custom set of inferences.
    [0020]
    20. Computer readable storage medium (325;330) having a method characterized in that it comprises: training a default template on a computing device (210) to customize a specific container template, wherein the default template is formed to from a plurality of weighted factors adjusted against a sample of users having common characteristics with the recipient, the common characteristics being selected from a group consisting of: common vocation and common interest, and the recipient-specific model is formed from the default model , which is modified using the recipient's historical behavior and feedback information; intercepting an item addressed to the recipient on the computing device, where the selected item from a group includes e-mail message, scheduled message, instant message, based message in web and social collaboration messaging; extract (110) a plurality of item resources associated with the item on the device computational; wherein the item features include an item characteristic selected from a group including: item issuer characteristic, item container characteristic; conversation feature, and attachment feature; retrieve (115) the container-specific model, wherein the container-specific model comprises the plurality of weighted factors associated with the plurality of extracted item resources; importance to the plurality of extracted item features, including forming a combination of the plurality of weighted factors by calculating an importance weight as a probability range of threshold values; generating a predicted item importance based on the combination of the plurality of weighted factors, where predicted item importance designates the item as important or unimportant; enable at least one application feature associated with the item for the container based on predicted item importance selected from a group including: an emphasizing feature to emphasize key content of the item, a preview feature to provide a quick preview of the item, and a rec notification bear to provide a temporary preview of the item; and periodically acquire the container and return behavior associated with the item for a predetermined period of time, to continue training on the default model to customize the specific container model.
    类似技术:
    公开号 | 公开日 | 专利标题
    BR112013012553B1|2021-06-08|method for screening electronic communications in a computer system, computing device, and computer-readable storage medium environment
    US10515107B2|2019-12-24|Systems and methods for processing and organizing electronic content
    US20120143806A1|2012-06-07|Electronic Communications Triage
    US9886664B2|2018-02-06|System and method of message thread management
    US20140229880A1|2014-08-14|Systems and methods for prioritizing notifications on mobile devices
    US20140059141A1|2014-02-27|Electronic messaging system utilizing social classification rules
    US20070179945A1|2007-08-02|Determining relevance of electronic content
    US20090150507A1|2009-06-11|System and method for prioritizing delivery of communications via different communication channels
    US20160269337A1|2016-09-15|Extended email functionality
    WO2018137668A1|2018-08-02|Personalized message priority classification
    US9929996B2|2018-03-27|Common email database for a plurality of users
    US10356031B2|2019-07-16|Prioritized communication inbox
    WO2016183257A1|2016-11-17|Communication item insights
    US10742581B2|2020-08-11|Summarization-based electronic message actions
    US10558643B2|2020-02-11|Notifications system for content collaborations
    US10650098B2|2020-05-12|Content analyzer and recommendation tool
    US11184317B1|2021-11-23|Simplified user interface for viewing the current status of message threads
    WO2016148814A1|2016-09-22|Extended email functionality
    同族专利:
    公开号 | 公开日
    RU2013125971A|2014-12-10|
    KR101843604B1|2018-03-29|
    BR112013012553A2|2016-08-30|
    US20120143798A1|2012-06-07|
    JP2014508980A|2014-04-10|
    AU2011338871B2|2016-06-16|
    WO2012078342A3|2012-08-02|
    CA2817230C|2019-02-12|
    EP2649535A4|2014-08-06|
    CA2817230A1|2012-06-14|
    EP2649535B1|2020-04-29|
    WO2012078342A2|2012-06-14|
    RU2600102C2|2016-10-20|
    EP2649535A2|2013-10-16|
    JP6246591B2|2017-12-13|
    AU2011338871A1|2013-05-30|
    KR20140012621A|2014-02-03|
    CN102567091B|2014-09-17|
    US8744979B2|2014-06-03|
    CN102567091A|2012-07-11|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    JP3300584B2|1994-11-24|2002-07-08|松下電器産業株式会社|Optimization adjustment method and optimization adjustment device|
    US6004015A|1994-11-24|1999-12-21|Matsushita Electric Industrial Co., Ltd.|Optimization adjusting method and optimization adjusting apparatus|
    US6460036B1|1994-11-29|2002-10-01|Pinpoint Incorporated|System and method for providing customized electronic newspapers and target advertisements|
    US5917489A|1997-01-31|1999-06-29|Microsoft Corporation|System and method for creating, editing, and distributing rules for processing electronic messages|
    US7117358B2|1997-07-24|2006-10-03|Tumbleweed Communications Corp.|Method and system for filtering communication|
    US6393465B2|1997-11-25|2002-05-21|Nixmail Corporation|Junk electronic mail detector and eliminator|
    US6891567B2|1998-06-26|2005-05-10|Fotonation Holdings, Llc|Camera messaging and advertisement system|
    US6424997B1|1999-01-27|2002-07-23|International Business Machines Corporation|Machine learning based electronic messaging system|
    US7120865B1|1999-07-30|2006-10-10|Microsoft Corporation|Methods for display, notification, and interaction with prioritized messages|
    US7000194B1|1999-09-22|2006-02-14|International Business Machines Corporation|Method and system for profiling users based on their relationships with content topics|
    US6865582B2|2000-01-03|2005-03-08|Bechtel Bwxt Idaho, Llc|Systems and methods for knowledge discovery in spatial data|
    US6691106B1|2000-05-23|2004-02-10|Intel Corporation|Profile driven instant web portal|
    US6832244B1|2000-09-21|2004-12-14|International Business Machines Corporation|Graphical e-mail content analyser and prioritizer including hierarchical email classification system in an email|
    KR100388254B1|2000-10-21|2003-06-25|비앤텍|Method Of Representing And Controling Email Using Diary Forms And System Thereof|
    EP1326189A3|2001-12-12|2005-08-17|Microsoft Corporation|Controls and displays for acquiring preferences, inspecting behaviour, and guiding the learning and decision policies of an adaptive communications prioritization and routing systems|
    US7222156B2|2001-01-25|2007-05-22|Microsoft Corporation|Integrating collaborative messaging into an electronic mail program|
    US6901398B1|2001-02-12|2005-05-31|Microsoft Corporation|System and method for constructing and personalizing a universal information classifier|
    US7165105B2|2001-07-16|2007-01-16|Netgenesis Corporation|System and method for logical view analysis and visualization of user behavior in a distributed computer network|
    WO2003017055A2|2001-08-15|2003-02-27|Visa International Service Association|Method and system for delivering multiple services electronically to customers via a centralized portal architecture|
    US7519589B2|2003-02-04|2009-04-14|Cataphora, Inc.|Method and apparatus for sociological data analysis|
    CN1332333C|2002-02-19|2007-08-15|波斯蒂尼公司|E-mail management services|
    US7162494B2|2002-05-29|2007-01-09|Sbc Technology Resources, Inc.|Method and system for distributed user profiling|
    WO2004030296A1|2002-09-30|2004-04-08|Corposoft Ltd.|Method and devices for prioritizing electronic messages|
    US7469280B2|2002-11-04|2008-12-23|Sun Microsystems, Inc.|Computer implemented system and method for predictive management of electronic messages|
    US7249162B2|2003-02-25|2007-07-24|Microsoft Corporation|Adaptive junk message filtering system|
    US7653879B1|2003-09-16|2010-01-26|Microsoft Corporation|User interface for context sensitive creation of electronic mail message handling rules|
    US7996470B2|2003-10-14|2011-08-09|At&T Intellectual Property I, L.P.|Processing rules for digital messages|
    US8566263B2|2003-11-28|2013-10-22|World Assets Consulting Ag, Llc|Adaptive computer-based personalities|
    US7454716B2|2003-12-22|2008-11-18|Microsoft Corporation|Clustering messages|
    US20050204009A1|2004-03-09|2005-09-15|Devapratim Hazarika|System, method and computer program product for prioritizing messages|
    US7818377B2|2004-05-24|2010-10-19|Microsoft Corporation|Extended message rule architecture|
    US7941491B2|2004-06-04|2011-05-10|Messagemind, Inc.|System and method for dynamic adaptive user-based prioritization and display of electronic messages|
    US8161122B2|2005-06-03|2012-04-17|Messagemind, Inc.|System and method of dynamically prioritized electronic mail graphical user interface, and measuring email productivity and collaboration trends|
    EP1767010B1|2004-06-15|2015-11-11|Tekelec Global, Inc.|Method, system, and computer program products for content-based screening of MMS messages|
    US20060031347A1|2004-06-17|2006-02-09|Pekka Sahi|Corporate email system|
    JP2006004307A|2004-06-21|2006-01-05|Hitachi Ltd|Business assessment support method|
    GB0422441D0|2004-10-08|2004-11-10|I Cd Publishing Uk Ltd|Processing electronic communications|
    US20060080393A1|2004-10-12|2006-04-13|Cardone Richard J|Method for using e-mail documents to create and update address lists|
    US7487214B2|2004-11-10|2009-02-03|Microsoft Corporation|Integrated electronic mail and instant messaging application|
    US8065369B2|2005-02-01|2011-11-22|Microsoft Corporation|People-centric view of email|
    US20060195467A1|2005-02-25|2006-08-31|Microsoft Corporation|Creation and composition of sets of items|
    US20060294191A1|2005-06-24|2006-12-28|Justin Marston|Providing context in an electronic messaging system|
    US7853485B2|2005-11-22|2010-12-14|Nec Laboratories America, Inc.|Methods and systems for utilizing content, dynamic patterns, and/or relational information for data analysis|
    US7894677B2|2006-02-09|2011-02-22|Microsoft Corporation|Reducing human overhead in text categorization|
    US8019632B2|2006-10-16|2011-09-13|Accenture Global Services Limited|System and method of integrating enterprise applications|
    WO2008134708A1|2007-04-30|2008-11-06|Etelemetry, Inc.|Method and system for activity monitoring and forecasting|
    US8068588B2|2007-06-26|2011-11-29|Microsoft Corporation|Unified rules for voice and messaging|
    US8230024B2|2007-06-28|2012-07-24|Microsoft Corporation|Delegating instant messaging sessions|
    US9275118B2|2007-07-25|2016-03-01|Yahoo! Inc.|Method and system for collecting and presenting historical communication data|
    US20090043621A1|2007-08-09|2009-02-12|David Kershaw|System and Method of Team Performance Management Software|
    WO2009065045A1|2007-11-14|2009-05-22|Qualcomm Incorporated|Methods and systems for determining a geographic user profile to determine suitability of targeted content messages based on the profile|
    US9203911B2|2007-11-14|2015-12-01|Qualcomm Incorporated|Method and system for using a cache miss state match indicator to determine user suitability of targeted content messages in a mobile environment|
    US20090150507A1|2007-12-07|2009-06-11|Yahoo! Inc.|System and method for prioritizing delivery of communications via different communication channels|
    US20090222298A1|2008-02-29|2009-09-03|International Business Machines Corporation|Data Mining Method for Automatic Creation of Organizational Charts|
    US8195588B2|2008-04-03|2012-06-05|At&T Intellectual Property I, L.P.|System and method for training a critical e-mail classifier using a plurality of base classifiers and N-grams|
    US8682819B2|2008-06-19|2014-03-25|Microsoft Corporation|Machine-based learning for automatically categorizing data on per-user basis|
    US8458153B2|2008-08-26|2013-06-04|Michael Pierce|Web-based services for querying and matching likes and dislikes of individuals|
    US20100153318A1|2008-11-19|2010-06-17|Massachusetts Institute Of Technology|Methods and systems for automatically summarizing semantic properties from documents with freeform textual annotations|
    CN101533366A|2009-03-09|2009-09-16|浪潮电子信息产业股份有限公司|Method for acquiring and analyzing performance data of server|
    US20110055264A1|2009-08-28|2011-03-03|Microsoft Corporation|Data mining organization communications|
    US9529864B2|2009-08-28|2016-12-27|Microsoft Technology Licensing, Llc|Data mining electronic communications|
    US8655647B2|2010-03-11|2014-02-18|Microsoft Corporation|N-gram selection for practical-sized language models|
    US8744979B2|2010-12-06|2014-06-03|Microsoft Corporation|Electronic communications triage using recipient's historical behavioral and feedback|US20110055264A1|2009-08-28|2011-03-03|Microsoft Corporation|Data mining organization communications|
    US9529864B2|2009-08-28|2016-12-27|Microsoft Technology Licensing, Llc|Data mining electronic communications|
    US8744979B2|2010-12-06|2014-06-03|Microsoft Corporation|Electronic communications triage using recipient's historical behavioral and feedback|
    US10366341B2|2011-05-11|2019-07-30|Oath Inc.|Mining email inboxes for suggesting actions|
    US10453030B2|2012-06-20|2019-10-22|Wendy H. Park|Ranking notifications based on rules|
    CN104584017B|2012-08-16|2019-02-15|金格输入输出有限公司|Method for being modeled to behavior and changes in health|
    US9146895B2|2012-09-26|2015-09-29|International Business Machines Corporation|Estimating the time until a reply email will be received using a recipient behavior model|
    CN104065628B|2013-03-22|2017-08-11|腾讯科技（深圳）有限公司|Conversation processing method and device|
    WO2015047323A1|2013-09-27|2015-04-02|Hewlett-Packard Development Company, L. P.|Notifying a user of critical emails via text messages|
    US20150142717A1|2013-11-19|2015-05-21|Microsoft Corporation|Providing reasons for classification predictions and suggestions|
    US10997183B2|2013-12-05|2021-05-04|LenovoPte. Ltd.|Determining trends for a user using contextual data|
    GB2521637A|2013-12-24|2015-07-01|Ibm|Messaging digest|
    US10546099B2|2014-10-15|2020-01-28|Brighterion, Inc.|Method of personalizing, individualizing, and automating the management of healthcare fraud-waste-abuse to unique individual healthcare providers|
    US10896421B2|2014-04-02|2021-01-19|Brighterion, Inc.|Smart retail analytics and commercial messaging|
    US9577867B2|2014-05-01|2017-02-21|International Business Machines Corporation|Determining a time before a post is viewed by a recipient|
    RU2608880C2|2014-05-22|2017-01-25|Общество С Ограниченной Ответственностью "Яндекс"|Electronic device and method of electronic message processing|
    US20150339673A1|2014-10-28|2015-11-26|Brighterion, Inc.|Method for detecting merchant data breacheswith a computer network server|
    US20160063502A1|2014-10-15|2016-03-03|Brighterion, Inc.|Method for improving operating profits with better automated decision making with artificial intelligence|
    US20150066771A1|2014-08-08|2015-03-05|Brighterion, Inc.|Fast access vectors in real-time behavioral profiling|
    US20160071017A1|2014-10-15|2016-03-10|Brighterion, Inc.|Method of operating artificial intelligence machines to improve predictive model training and performance|
    US11080709B2|2014-10-15|2021-08-03|Brighterion, Inc.|Method of reducing financial losses in multiple payment channels upon a recognition of fraud first appearing in any one payment channel|
    US9280661B2|2014-08-08|2016-03-08|Brighterion, Inc.|System administrator behavior analysis|
    US20160055427A1|2014-10-15|2016-02-25|Brighterion, Inc.|Method for providing data science, artificial intelligence and machine learning as-a-service|
    US10671915B2|2015-07-31|2020-06-02|Brighterion, Inc.|Method for calling for preemptive maintenance and for equipment failure prevention|
    US20160078367A1|2014-10-15|2016-03-17|Brighterion, Inc.|Data clean-up method for improving predictive model training|
    US10942946B2|2016-09-26|2021-03-09|Splunk, Inc.|Automatic triage model execution in machine data driven monitoring automation apparatus|
    US10290001B2|2014-10-28|2019-05-14|Brighterion, Inc.|Data breach detection|
    US10504029B2|2015-06-30|2019-12-10|Microsoft Technology Licensing, Llc|Personalized predictive models|
    KR101704651B1|2015-07-06|2017-02-08|주식회사 이팝콘|Method for providing message service rule based on reception analysis and method for sending message using message service rule|
    KR101688829B1|2015-07-24|2016-12-22|삼성에스디에스 주식회사|Method and apparatus for providing documents reflecting user pattern|
    AU2015224398A1|2015-09-08|2017-03-23|Canon Kabushiki Kaisha|A method for presenting notifications when annotations are received from a remote device|
    WO2017048730A1|2015-09-14|2017-03-23|Cogito Corporation|Systems and methods for identifying human emotions and/or mental health states based on analyses of audio inputs and/or behavioral data collected from computing devices|
    US9923853B2|2015-10-05|2018-03-20|Quest Software Inc.|Folders that employ dynamic user training rules to organize content|
    US10574600B1|2016-03-25|2020-02-25|Amazon Technologies, Inc.|Electronic mailbox for online and offline activities|
    US10749833B2|2016-07-07|2020-08-18|Ringcentral, Inc.|Messaging system having send-recommendation functionality|
    JP2019079224A|2017-10-24|2019-05-23|富士ゼロックス株式会社|Information processing device and information processing program|
    US10579632B2|2017-12-18|2020-03-03|Microsoft Technology Licensing, Llc|Personalized content authoring driven by recommendations|
    US11176472B2|2018-05-22|2021-11-16|International Business Machines Corporation|Chat delta prediction and cognitive opportunity system|
    US11159476B1|2019-01-31|2021-10-26|Slack Technologies, Llc|Methods and apparatuses for managing data integration between an external email resource and a group-based communication system|
    WO2021192214A1|2020-03-27|2021-09-30|日本電気株式会社|Stress management device, stress management method, and computer-readable recording medium|
    CN112565803A|2020-11-30|2021-03-26|北京达佳互联信息技术有限公司|Comment area message processing method and device and computer storage medium|
    CN113421054A|2021-06-11|2021-09-21|荣耀终端有限公司|Information management method, electronic device, and storage medium|
    法律状态:
    2018-02-27| B25A| Requested transfer of rights approved|Owner name: MICROSOFT TECHNOLOGY LICENSING, LLC (US) |
    2018-12-18| B06F| Objections, documents and/or translations needed after an examination request according [chapter 6.6 patent gazette]|
    2019-10-01| B06U| Preliminary requirement: requests with searches performed by other patent offices: procedure suspended [chapter 6.21 patent gazette]|
    2020-11-17| B06A| Patent application procedure suspended [chapter 6.1 patent gazette]|
    2021-03-30| B09A| Decision: intention to grant [chapter 9.1 patent gazette]|
    2021-06-08| B16A| Patent or certificate of addition of invention granted [chapter 16.1 patent gazette]|Free format text: PRAZO DE VALIDADE: 20 (VINTE) ANOS CONTADOS A PARTIR DE 20/11/2011, OBSERVADAS AS CONDICOES LEGAIS. |
    优先权:
    申请号 | 申请日 | 专利标题
    US12/961,180|2010-12-06|
    US12/961,180|US8744979B2|2010-12-06|2010-12-06|Electronic communications triage using recipient's historical behavioral and feedback|
    PCT/US2011/061571|WO2012078342A2|2010-12-06|2011-11-20|Electronic communications triage|
    [返回顶部]