Abstract: A computer-implemented method for detecting and monitoring bias in an application includes index training data and obtaining a plurality of correlation values of one or more features in the indexed training data with a target variable. For each of the one or more features, a first value and a favorable result and a second value along with the unfavorable result is calculated. An absolute value of a difference between the calculated first value and the calculated second value is calculated. A total sum of the calculated absolute value of the plurality of correlation values of the one of the one or more features is calculated.

Abstract: A computer-implemented method, system and computer program product for improving the quality of subsequent searches. A request is received from a user to perform a search using the term(s) provided by the user. A search is then performed using the provided term(s). A representation of the search result is depicted as an area (e.g., bubble) with terms of interest displayed outside the area. The area represents a set of documents that were found in the performed search and the terms of interest represent terms that may be of interest to the user based on this set of documents. These terms of interest may be selected by the user to adjust the search result. In this manner, the user is presented with an intuitive view as to how different terms or different searches will have an impact on the search results thereby improving the quality of subsequent searches.

Abstract: A method and indexing system indexes the content of a body of documents into a content index, and the metadata of the documents into a metadata index which is a parallel index to the content index. The metadata is copied into a data store that is easily accessible by the indexing system and is stored in native form. The indexing system can dynamically re-index the metadata from the native metadata in the data store to produce a new metadata index which is used to replace the original metadata index. Search queries received by a search engine associated with the indexing system are applied to both the content and metadata index and the results are merged for return.

Abstract: In an approach to AI explaining for natural language processing, responsive to receiving an input text for a machine learning model, an output is generated from the machine learning model. A plurality of alteration techniques are applied to the input text to generate one or more alternate outputs, where each alternate output corresponds to an alteration technique. A variation rate of the alternate output is calculated for each alteration technique. A preferred technique of generating neighboring data of the input text is generated based on a comparison of the variation rate of the alternate output for each alteration technique.

Abstract: A computer-implemented method, system and computer program product for improving the quality of subsequent searches. A request is received from a user to perform a search using the term(s) provided by the user. A search is then performed using the provided term(s). A representation of the search result is depicted as an area (e.g., bubble) with terms of interest displayed outside the area. The area represents a set of documents that were found in the performed search and the terms of interest represent terms that may be of interest to the user based on this set of documents. These terms of interest may be selected by the user to adjust the search result. In this manner, the user is presented with an intuitive view as to how different terms or different searches will have an impact on the search results thereby improving the quality of subsequent searches.

Abstract: A computer-implemented method, system and computer program product for improving the quality of subsequent searches. A request is received from a user to perform a search using the term(s) provided by the user. A search is then performed using the provided term(s). A representation of the search result is depicted as an area (e.g., bubble) with terms of interest displayed outside the area. The area represents a set of documents that were found in the performed search and the terms of interest represent terms that may be of interest to the user based on this set of documents. These terms of interest may be selected by the user to adjust the search result. In this manner, the user is presented with an intuitive view as to how different terms or different searches will have an impact on the search results thereby improving the quality of subsequent searches.

Abstract: A computer-implemented method, system and computer program product for improving the quality of subsequent searches. A request is received from a user to perform a search using the term(s) provided by the user. A search is then performed using the provided term(s). A representation of the search result is depicted as an area (e.g., bubble) with terms of interest displayed outside the area. The area represents a set of documents that were found in the performed search and the terms of interest represent terms that may be of interest to the user based on this set of documents. These terms of interest may be selected by the user to adjust the search result. In this manner, the user is presented with an intuitive view as to how different terms or different searches will have an impact on the search results thereby improving the quality of subsequent searches.

Abstract: A computer-implemented method, system and computer program product for improving the quality of subsequent searches. A request is received from a user to perform a search using the term(s) provided by the user. A search is then performed using the provided term(s). A representation of the search result is depicted as an area (e.g., bubble) with terms of interest displayed outside the area. The area represents a set of documents that were found in the performed search and the terms of interest represent terms that may be of interest to the user based on this set of documents. These terms of interest may be selected by the user to adjust the search result. In this manner, the user is presented with an intuitive view as to how different terms or different searches will have an impact on the search results thereby improving the quality of subsequent searches.

Abstract: A method for recommending a plurality of alternate search keywords is provided. The method may include executing a first search query using a plurality of user-entered search text. The method may also include identifying a highest contribution keyword to a plurality of search results of the executed first search query. The method may further include identifying a highest correlation alternate keyword to the identified highest contribution keyword. The method may also include creating an alternate keyword group by replacing the identified highest contribution keyword with the identified highest correlation alternate keyword. The method may further include executing a second search query using the created alternate keyword group. The method may also include displaying the plurality of search results associated with the executed first search query with a plurality of statistics associated with the executed second search query.

Abstract: The present invention provides a method, system, and computer program product of managing and displaying online messages along timelines. In an embodiment, the present invention includes receiving a plurality of online messages from an online website, determining sets of similar online messages among the plurality of online messages, and displaying on a computer display each of the sets of similar online messages along a graphical representation of a timeline corresponding to the each of the sets of similar online messages, resulting in displayed timelines and displayed online messages. In a further embodiment, the present invention includes in response to receiving a new online message from the online website, determining each set among the sets of similar online messages to which the new online message is similar, and displaying on the computer display the new online message, resulting in a displayed new online message.

Abstract: In an approach to AI explaining for natural language processing, responsive to receiving an input text for a machine learning model, an output is generated from the machine learning model. A plurality of alteration techniques are applied to the input text to generate one or more alternate outputs, where each alternate output corresponds to an alteration technique. A variation rate of the alternate output is calculated for each alteration technique. A preferred technique of generating neighboring data of the input text is generated based on a comparison of the variation rate of the alternate output for each alteration technique.

Abstract: A computer-implemented method for detecting and monitoring bias in an application includes index training data and obtaining a plurality of correlation values of one or more features in the indexed training data with a target variable. For each of the one or more features, a first value and a favorable result and a second value along with the unfavorable result is calculated. An absolute value of a difference between the calculated first value and the calculated second value is calculated. A total sum of the calculated absolute value of the plurality of correlation values of the one of the one or more features is calculated.

Abstract: A method, computer system, and computer program product for segment differential-based document text-index modeling are provided. The embodiment may include receiving, by a processor, a document with a valid document ID and version ID tuple. The embodiment may also include determining the received document is a new version of a previously stored document and consequently multiplexing versions of the document into a single indexed document. The embodiment may further include segmenting the received document and building a token vector. The embodiment may also include calculating a difference between the received new version of the document and the previously stored document using information obtained from the segmentation. The embodiment may further include in response to the calculated difference being below a pre-configured threshold value, discarding the received new version.

Abstract: A first character string in a logic, wherein the logic is a rule which analyzes input text and extracts additional information used for the text analysis, used for the text analysis with a second character string which is different form the first character string is replaced using a mapping between the first character string and the second character string. The first character string being a character or a combination of characters in a logic. A first character string is searched for in a target document for the text analysis and the first character string, if found, is replaced with the second character string, using the mapping. The logic is sent after the replacement to a server computer. The target document after the replacement is sent to the server computer. A text analysis result for the target document after the replacement is received form the server computer.

Abstract: An apparatus includes a receiving unit operable to receive a plurality of outputs from a plurality of hosts, a comparing unit operable to compare the plurality of outputs, and a display processing unit operable to generate a display in which portions of the plurality of outputs are distinguished based on a comparison result of the comparing unit.

Abstract: An apparatus includes a receiving unit operable to receive a plurality of outputs from a plurality of hosts, a comparing unit operable to compare the plurality of outputs, and a display processing unit operable to generate a display in which portions of the plurality of outputs are distinguished based on a comparison result of the comparing unit.

Abstract: A method, computer system, and computer program product for segment differential-based document text-index modeling are provided. The embodiment may include receiving, by a processor, a document with a valid document ID and version ID tuple. The embodiment may also include determining the received document is a new version of a previously stored document and consequently multiplexing versions of the document into a single indexed document. The embodiment may further include segmenting the received document and building a token vector. The embodiment may also include calculating a difference between the received new version of the document and the previously stored document using information obtained from the segmentation. The embodiment may further include in response to the calculated difference being below a pre-configured threshold value, discarding the received new version.

Abstract: The present invention provides a method, system, and computer program product of managing and displaying online messages along timelines. In an embodiment, the present invention includes receiving a plurality of online messages from an online website, determining sets of similar online messages among the plurality of online messages, and displaying on a computer display each of the sets of similar online messages along a graphical representation of a timeline corresponding to the each of the sets of similar online messages, resulting in displayed timelines and displayed online messages. In a further embodiment, the present invention includes in response to receiving a new online message from the online website, determining each set among the sets of similar online messages to which the new online message is similar, and displaying on the computer display the new online message, resulting in a displayed new online message.

Abstract: Embodiments are directed to a computer implemented method of improving queries. The method includes performing a query on a document set, wherein documents that satisfy the query comprise a first subset of the document set and documents that do not satisfy the query comprise a complement subset. Thereafter, calculating a correlation for all terms in the complement subset, with respect to the complement subset. Thereafter determining a set of top correlated terms. A set of bottom correlated terms is determined. The set of top correlated terms and bottom correlated terms are presented to generate an improved query.

Abstract: The present invention provides a method, system, and computer program product of managing and displaying online messages along timelines. In an embodiment, the present invention includes receiving a plurality of online messages from an online website, determining sets of similar online messages among the plurality of online messages, and displaying on a computer display each of the sets of similar online messages along a graphical representation of a timeline corresponding to the each of the sets of similar online messages, resulting in displayed timelines and displayed online messages. In a further embodiment, the present invention includes in response to receiving a new online message from the online website, determining each set among the sets of similar online messages to which the new online message is similar, and displaying on the computer display the new online message, resulting in a displayed new online message.