Abstract: A method for training a machine learning model for open domain question answering includes receiving trained classifiers for question answering. The received trained classifiers are used to generate a set of candidate answers to a question. Second trained classifiers are used for scoring each of the candidate answers. The scoring indicates a measure of how well each candidate answer answers the question. Using the second trained classifiers for scoring each of the candidate answers includes comparing each candidate answer to a first ground truth corresponding to the question. A set of top-scoring candidate answers is presented to a human operator who marks each as correct or incorrect. The correct candidate answers are treated as additional ground truths for further training the first trained classifiers.

Abstract: A method for training a machine learning model for open domain question answering includes receiving trained classifiers for question answering. The received trained classifiers are used to generate a set of candidate answers to a question. Second trained classifiers are used for scoring each of the candidate answers. The scoring indicates a measure of how well each candidate answer answers the question. Using the second trained classifiers for scoring each of the candidate answers includes comparing each candidate answer to a first ground truth corresponding to the question. A set of top-scoring candidate answers is presented to a human operator who marks each as correct or incorrect. The correct candidate answers are treated as additional ground truths for further training the first trained classifiers.

Abstract: Embodiments of the present invention provide a method, system and computer program product for displaying search results for weighted, multi-term content searches. In an embodiment of the invention, a method for displaying search results for weighted, multi-term content searches is provided. The method includes specifying different weighted search terms for a content search and performing the content search for both content and content meta-data in a search engine. The method also includes retrieving search results for the content search and computing a relevance for each of the weighted search terms. Finally, the method includes presenting both the search results and also a relevance indicator for each computed relevance for each of the weighted search terms found in connection with each of the search results in a user interface to the search engine.

Abstract: A computer-implemented method of selecting a real time collaboration application can include associating each of a plurality of real time collaboration applications with a plurality of performance statistics, receiving a request for a real time collaboration session from a personal productivity application, and identifying a data type to be shared through one of the plurality of real time collaboration applications. One of the plurality of real time collaboration applications can be selected according to the data type to be shared and the plurality of performance statistics. The selected real time collaboration application can be output.

Abstract: Embodiments of the present invention provide a method, system and computer program product for game determined tag relevance for social bookmarking. In one embodiment of the invention, a method for the game determination of tag relevance for social bookmarking can be provided. The method can include rendering a game user interface in connection with content in a content browser, providing a list of tags previously associated with the content in the game user interface for a game participant and receiving relevance values for the tags through the game user interface provided by the game participant for each of the tags, comparing the relevance values to relevance values provided by others to generate a score and presenting the score in the game user interface, and applying the received relevance values to the tags to improve quality for each of the tags in association with the content.

Abstract: A computer-implemented method of determining a common social context can include detecting a collaboration between a plurality of participants. The method can include identifying a plurality of common social contexts and determining a probability that the collaboration belongs to each of the plurality of common social contexts. At least one probability and corresponding common social context can be output.

Abstract: It may be detected that a user of a communication application engages in an activity on a website. User preferences of the user then may be applied. If it is allowed by the user preferences, the fact that the user has engaged in the activity may be communicated to a process running on a computing device of at least one other user of the communication application. A list of one or more member identifiers may be generated on a computer display of the computing device. If allowed by user preferences of the user, a notification indicative of the recent activity engaged in by the member is generated on the computer display such that the notification is associated with at least one member identifier of the one or more member identifiers corresponding to the user.

Abstract: Embodiments of the present invention provide a method, system and computer program product for displaying search results for weighted, multi-term content searches. In an embodiment of the invention, a method for displaying search results for weighted, multi-term content searches is provided. The method includes specifying different weighted search terms for a content search and performing the content search for both content and content meta-data in a search engine. The method also includes retrieving search results for the content search and computing a relevance for each of the weighted search terms. Finally, the method includes presenting both the search results and also a relevance indicator for each computed relevance for each of the weighted search terms found in connection with each of the search results in a user interface to the search engine.

Abstract: A content searching data processing system can be configured for specifying weighted search terms for a search engine. The system can include a search engine executing in a host server. The search engine can be coupled to a search index and can be configured for communicative coupling to different content browsers executing in respective clients over a computer communications network. Finally, the system can include weighted search term logic coupled to the search engine. The logic can include program code enabled to render a search term entry user interface in which variable weights are specified for corresponding search terms, to assign the variable weights to the corresponding search terms and to issue a search to the search engine with the search terms and variable weights and to return results of the search to a requesting one of the content browsers.

Abstract: Embodiments of the present invention provide a method, system and computer program product for game determined tag relevance for social bookmarking. In one embodiment of the invention, a method for the game determination of tag relevance for social bookmarking can be provided. The method can include rendering a game user interface in connection with content in a content browser, providing a list of tags previously associated with the content in the game user interface for a game participant and receiving relevance values for the tags through the game user interface provided by the game participant for each of the tags, comparing the relevance values to relevance values provided by others to generate a score and presenting the score in the game user interface, and applying the received relevance values to the tags to improve quality for each of the tags in association with the content.

Abstract: A method of determining a common social context can include detecting a collaboration between a plurality of participants. The method can include identifying a plurality of common social contexts and determining a probability that the collaboration belongs to each of the plurality of common social contexts. At least one probability and corresponding common social context can be output.

Abstract: A computer-implemented method of selecting a real time collaboration application can include associating each of a plurality of real time collaboration applications with a plurality of performance statistics, receiving a request for a real time collaboration session from a personal productivity application, and identifying a data type to be shared through one of the plurality of real time collaboration applications. One of the plurality of real time collaboration applications can be selected according to the data type to be shared and the plurality of performance statistics. The selected real time collaboration application can be output.

Abstract: An improved solution for managing a meeting. In particular, a meeting document is constructed in a hierarchical manner. Before the meeting, the meeting document can assist in planning the meeting by including a set of agenda items for the meeting and/or a list of participants. During the meeting, the meeting document can be used to conduct the meeting, generate minutes for the meeting, and/or index a log of the meeting. After the meeting, the meeting document can be stored for future reference.

Abstract: Described are a system and method of communicating with a user of a computing device over a network. Communications occur with a body-less electronic mail messages. The conveyed message is carried by the subject line of the body-less electronic mail message. The subject line records the temporal sequence of one or more text messages representing a chat-like email conversation between users. Each received body-less electronic mail messages appears on a display screen as a line item in a mailbox view. In one embodiment, the entire contents of the subject line are displayed in a drop-down list when a cursor is positioned over the subject column of the line item. In another embodiment, the user scrolls horizontally through contents of the subject line by positioning the cursor over a scroll bar arrow that appears at one end of the subject line column.

Abstract: Method, system and program product for conducting an on-line meeting. An on-line meeting participant application is used to enumerate the participants, including identifying each of the individual participants to individual on-line meeting sites, and publishing an agenda to enumerated participants. Individual participant inputs are collected from the individual participants and published in accordance with participant attributes.