Abstract: Disentanglement of chat utterances is provided. An analysis of the linguistic collocations and the keywords of the multiple chat utterances and amount of contribution by respective chat users of the plurality of chat users to the multiple chat utterances is performed to determine a level of drift of the linguistic collocations, the keywords, and respective chat users over a course of the multiple chat utterances. Chat utterance entanglement of prior chat utterances is determined using determined level of drift based on the analysis by inferring keyword usage over time and how these keywords are related over the course of the multiple chat utterances. The prior chat utterances related to a particular topic are disentangled by removing certain chat utterances that have a statistically significant level of drift from that particular topic. Removed chat utterances are arranged as a new chat discourse related to a different topic in the chat conversation.

Abstract: Disclosed are techniques to model when authoritative responses are expected to be received in real-time chat dialogues to facilitate monitoring a live real-time chat and connect a participant to provide an authoritative answer at an appropriate time during the live real-time chat. Historical real-time chat dialogues are used to generate a model for when authoritative responses are expected in real-time chats. A live real-time chat is monitored for an ongoing discussion without an authoritative response, and a response latency window of the model for the live real-time chat is adjusted based on the ongoing discussion in the live real-time chat. In some embodiments, upon entering an optimal window of the ongoing discussion of the live real-time chat indicated by the model where it is determined that no authoritative response is yet received, a new participant is connected to the live real-time chat to provide an authoritative response.

Abstract: A discourse pattern is generated by analyzing a set of natural language interactions. A sensor activation plan corresponding to the discourse pattern is determined, the sensor activation plan comprising a data collection parameter corresponding to a first sensor. Within a first natural language interaction, a first pattern having above a threshold similarity to the discourse pattern is detected. Responsive to the detecting, a configuration of the sensor is adjusted according to the sensor activation plan.

Abstract: Bystander effect management can include determining a likelihood of a bystander effect on at least one chat participant engaging in a chat discourse over an electronic communications network. The likelihood can be determined by a bystander effect perception (BEP) model generated based on collaboration interaction metrics derived from prior discourses conducted over the electronic communications network by a chat group. A bystander effect reaction (BER) can be generated in response to determining that the likelihood of the bystander effect exceeds a predetermined threshold, the BER being determined based on the BEP model to be more likely than not to enhance collaboration among at least some chat participants engaging in the chat discourse. A notification based on the BER can be conveyed to one or more chat participants engaging in the chat discourse over the electronic communications network.

Abstract: An embodiment of a summarization application divides collected conversation data into media and text components. The application implements respective machine learning mechanisms to enhance modeling operations of the text and media components to identify key elements from the conversation. The application generates a headline banner from a group of key elements based on an analysis involving first predetermined criteria. The application also combines additional key elements to the group of key elements to form a second group of key elements. The application generates a summary from the second group of key elements based on a second analysis involving second predetermined criteria. The application presents, via a display, the headline banner according to a first output of the first key element analysis and the summary according to a second output of the second key element analysis.

Abstract: A computing device that receives a real-time chat discourse. The computing device analyses the real-time chat discourse by consecutively applying a topic analysis technique, a corpus linguistics technique and a cosine similarity technique. The computing device derives a discourse decision forking component (DDFC) based on comparing the analyzed real-time chat discourse to a similarity threshold value and determines one or more discourse forks using the DDFC.

Abstract: A computer-implemented method is provided. The embodiments include evaluating, by one or more processors, a specimen chart relative to a chart erratum model that has features mapped to an optimum state for a first chart type. The method also includes generating a first risk score for a first sample feature of the specimen chart. The first risk score may include a delta from the optimum state. The method also includes refactoring the specimen chart to mitigate the first risk score of the first sample feature.

Abstract: A computing device that receives a real-time chat discourse. The computing device analyses the real-time chat discourse by consecutively applying a topic analysis technique, a corpus linguistics technique and a cosine similarity technique. The computing device derives a discourse decision forking component (DDFC) based on comparing the analyzed real-time chat discourse to a similarity threshold value and determines one or more discourse forks using the DDFC.

Abstract: In an approach for collaboration space identification using sensory alignment, a processor collects scheduling information and expected device usage from historical log files for a forthcoming time period for a business. A processor determines historical and expected energy consumption from the scheduling information and expected device usage for the forthcoming time period. A processor determines an expected power usage based on the historical and expected energy consumption for the forthcoming time period. A processor monitors net incoming electricity to compare against the expected power usage for the forthcoming time period. A processor determines a level of significance between each location of the business and each application expected to be used in the forthcoming time period. A processor builds a model based on each level of significance determined that determines a probability of whether a respective location is optimal for collaboration.

Abstract: In response to determining that a native language of a first user is different from a target language of a message to be transmitted by the first user to a second user, a translation model based on a plurality of language efficacies of the first user is created. An optimal action associated with a translation of the message from the native language to the target language is determined based on the created model and a language efficacy of the first user in the native language. The determined optimal action is performed. The message translation comprising the performed optimal action is transmitted to the second user.

Abstract: A set of baseline treatment information is received. From the set of baseline treatment information, a set of incidental information is determined. Based on the set of baseline treatment information and the set of incidental information, a treatment mitigation area is generated. Based on the set of baseline treatment information, the set of incidental information, and the treatment mitigation area, a treatment mitigation model is created. An affected individual is identified, based on the treatment mitigation model. The affected individual is notified of one or more treatment mitigation recommendations.

Abstract: Digital photograph filters and associated settings are recommended by an image classification model. A plurality of images and metadata associated with each of the plurality of images are received. Human interaction scores for each of the images are received. Training data is generated classifying the images using the associated metadata and human interaction scores. The training data is used to derive an image classification model for predicting a human interaction score for an image having an unknown human interaction score. At least one recommended image capture setting using the image classification model is determined in response to determining that a user is preparing to capture an image with a device. The recommended image capture setting is displayed on a display of the device.

Abstract: At least one target in a speech of a first speaker during an electronic conference is identified. A closest vector between the identified target and a group of possible matches for the target is determined. A most similar match for the identified target is determined based on a current usage context for the identified target, a history of stored information associated with the first speaker and the identified target, and the determined closest vector. The most similar match to a set of participants of the electronic conference is displayed.

Abstract: Communication between a plurality of users on one or more media platforms is analyzed. A model of the communication that reflects trends of the communication is generated. Using the model, it is determined that a current communication between a first and second user of the media platform is relatively less efficient than a potential communication between the first user and a third user of the media platform. In response to determining that the current communication is relatively less efficient than the potential communication, the current communication is directed to the third person.

Abstract: A method, computer system, and computer program product for collaborative document relations modeling are provided. The embodiment may include parsing, by a processor, a document corpus utilizing topic modeling and distance analysis techniques. The embodiment may also include deriving a collaborative document relational model to combine the results of the parsing into a matrix. The embodiment may further include grouping content of the parsed document corpus through document clustering utilizing the generated collaborative document relational model. The embodiment may also include displaying the grouped content as document clusters in a graphical user interface of a document management application.

Abstract: A system determines if a user is hesitating when inputting a command based on a time delay. The system determines a severity value based on the hesitancy, and further determines whether or not to prompt the user to validate the command based (at least in part) on the severity value.

Abstract: A method, system, and computer program product for creating step-based answers from multimedia content are provided. The method receives a search query including a set of search terms. A set of question characteristics is determined for the search query. The set of question characteristics are based in part on the set of search terms. The method generates a set of steps for an answer to the search query. The set of steps are generated based in part on the set of question characteristics. The method identifies media content relevant to the set of steps and segments the media content to generate a set of media segments. Each media segment of the set of media segments represent a portion of the media content. The method combines the set of media segments based on the set of steps to generate a media answer to the search query.

Abstract: An embodiment of a summarization application divides collected conversation data into media and text components. The application implements respective machine learning mechanisms to enhance modeling operations of the text and media components to identify key elements from the conversation. The application generates a headline banner from a group of key elements based on an analysis involving first predetermined criteria. The application also combines additional key elements to the group of key elements to form a second group of key elements. The application generates a summary from the second group of key elements based on a second analysis involving second predetermined criteria. The application presents, via a display, the headline banner according to a first output of the first key element analysis and the summary according to a second output of the second key element analysis.

Abstract: According to one embodiment, a method, computer system, and computer program product for detecting and repairing gaps within a call is provided. The present invention may include requesting a personal discourse gap recovery model (PDGRM) for participants in a call, where the PDGRM may be a machine learning model that models a user's speech patterns based on word collocation, dictionary, and speaking rate. The present invention may further comprise detecting one or more gaps in the call based on network connectivity, inserting an audio delay into the call, constructing repaired dialogue to fill at least one of the one or more gaps in the call based on the PDGRM, and inserting the repaired dialogue into the call.

Abstract: Aspects of the present invention disclose a method, computer program product, and system for generating messages in a real-time chat discourse. The method includes one or more processors identifying user profile data associated with a user engaged in a real-time chat discourse with a chatbot. The method further includes one or more processors analyzing the real-time chat discourse between the user and the chatbot. The method further includes one or more processors determining a baseline emotive level for the real-time chat discourse based on the user profile data and the analysis of the real-time chat discourse. The method further includes one or more processors determining real-time emotion information of messages of the user in the real-time chat discourse. The method further includes one or more processors determining whether the real-time emotion information deviates from the determined baseline emotive level.