Abstract: Embodiments of the present invention provide a method comprising receiving a task set comprising multiple tasks, receiving operational information identifying one or more operating characteristics of multiple drones, and obtaining an initial heuristic ordering of the multiple tasks based on the operational information and the climate information. Each task has a corresponding task location. The method further comprises scheduling the multiple tasks to obtain a final ordering of the multiple tasks. The final ordering represents an order in which the multiple tasks are scheduled, and the final ordering may be different from the initial heuristic ordering.

Abstract: A system includes a machine learning module configured to train a location prediction model for an information campaign, a front-end server configured to receive and process information requests, and a prediction unit. During the information campaign, the prediction unit is configured to use the location prediction model to predict a conversion probability for any particular mobile device associated with a qualified information request received during any respective time unit. The conversion probability corresponds to a predicted probability of the particular mobile device having at least one location event at any of one or more POIs during a particular time frame. The front-end server is further configured to determine a respective target number of conversions to be achieved by the information campaign during the respective time unit, and to determine a response to the particular information request based at least in part on the conversion probability and on the respective target number of conversions.

Abstract: A method and system of automatically identifying topics of a conversation are provided. An electronic data package comprising a sequence of utterances between conversation entities is received by a computing device. Each utterance is classified to a corresponding social action. One or more utterances in the sequence are grouped into a segment based on a deep learning model. A similarity of topics between adjacent segments is determined. Upon determining that the similarity is above a predetermined threshold, the adjacent segments are grouped together. A transcript of the conversation including the grouping of the adjacent segments is stored in a memory.

Abstract: A system includes a machine learning module configured to train a location prediction model using features constructed from mobile device data with time stamps in a training time period, and labels extracted from mobile device data with time stamps in a training time frame. The system further includes a prediction module configured apply the prediction model to a feature set constructed using mobile device data associated with a mobile device with time stamps in a prediction time period to obtain a prediction result corresponding to the mobile device. The system further includes a calibration module configured to obtain a calibration model corresponding to an information campaign, and a calibrated prediction module configured to apply the calibration model to the prediction result to obtain a calibrated probability for the mobile device to have at least one location event at any of one or more locations associated with the information campaign during a prediction time frame.

Abstract: A computer-implemented method according to one embodiment includes determining a starting folder within a file system, computing, for each child folder of the starting folder, a similarity metric indicating a level of similarity to a file, selecting two child folders of the starting folder having greatest similarity metrics, comparing a difference between the greatest similarity metrics of the two child folders to a predetermined threshold, and conditionally selecting the starting folder as a recommended folder to which the file is saved, based on the comparing.

Abstract: One embodiment provides a method comprising receiving historic peer deals relating to at least one service, and a baseline and cost percentage estimation for each service. Historic peer cost data for each service is clustered to form at least one cluster. Each cluster includes similar unit costs, and has an assigned label. A classification model is trained based on each baseline received, each cost percentage estimation received, and each assigned label. For each assigned label, a corresponding probability distribution is computed based on the classification model. For each service of a new client solution, an assigned label for the service is predicted based on the classification model, and, based on a probability distribution corresponding to the assigned label predicted, transforming an initial range of historic peer cost data relating to the service into a narrower range for use in estimating a cost of the service with improved accuracy.

Abstract: A computer-implemented method according to one embodiment includes determining a starting folder within a file system, computing, for each child folder of the starting folder, a similarity metric indicating a level of similarity to a file, selecting two child folders of the starting folder having greatest similarity metrics, comparing a difference between the greatest similarity metrics of the two child folders to a predetermined threshold, and conditionally selecting the starting folder as a recommended folder to which the file is saved, based on the comparing.

Abstract: A computer-implemented method according to one embodiment includes creating a profile for a user, the profile including one or more default aspects created automatically by a system and one or more custom aspects created in response to textual input by the user, comparing event data to the profile, and providing a recommendation to the user, based on the comparing.

Abstract: One embodiment provides a method for augmenting missing values in historical or market data for deals. The method comprises receiving information relating to a set of deals. For any service included in one or more deals of the set of deals but not included in one or more other deals of the set of deals, the method further comprises augmenting, for any or all of the one or more other deals that does not include the service, one or more missing values for the service with one or more recommended values based on a recommendation algorithm. The service may be at any service level of a hierarchy of services.

Abstract: An example operation may include one or more of identifying a first conferred asset exchange request from a first user account and a second conferred asset exchange request from a second user account which are capable of being used to settle each other, requesting an intermediary blockchain to perform a conferred asset settlement transaction for the first and second conferred asset exchange requests, determining that first conferred assets of the first user account and second conferred assets of the second user account have been transferred to temporary intermediary trading addresses, respectively, and releasing the first conferred assets to the second user account and the second loyalty assets to the first user account, in response to the determining.

Abstract: One embodiment provides a method for predicting a next action in a conversation system that includes obtaining, by a processor, information from conversation logs and a conversation design. The processor further creates a dialog graph based on the conversation design. Weights and attributes for edges in the dialog graph are determined based on the information from the conversation logs and adding user input and external context information to an edge attributes set. An unrecognized user input is analyzed and a next action is predicted based on dialog nodes in the dialog graph and historical paths. A guiding conversation response is generated based on the predicted next action.

Abstract: A method and system of evaluating a user experience (UX) design are provided. A UX design is received. All objects that are identified to be part of a background of the input UI screen are removed to create a filtered input UI screen. The input UI screen is assigned to a cluster. A target UI screen of the input screen is determined and its background removed, to create a filtered target UI cluster. The target UI screen is assigned to a cluster. The filtered input UI screen is used as an input to a deep learning model to predict a target UI cluster. The predicted target UI cluster is compared to the filtered target UI cluster based on the clustering. Upon determining that the filtered target UI cluster is similar to the target UI screen, the UX design is classified as being successful.

Abstract: In one general aspect, a computer-implemented method includes identifying current choices with different verbosity levels for a current turn in a conversation; normalizing multi-dimensional verbosity vectors for each of the current choices to obtain a normalized value for each of the current choices; determining a state definition for the current turn in the conversation, utilizing the normalized values for each of the current choices; providing the state definition for the current turn in the conversation and the normalized values for each of the current choices to a trained reinforcement learning module; receiving, from the trained reinforcement learning module, a score associated with each of the current choices for the current turn in the conversation; and selecting one of the current choices to be entered for the current turn in the conversation, based on the score associated with each of the current choices for the current turn in the conversation.

Abstract: A method, a computer program product, and a computer system counterbalance a developed bias of user reviews. The method includes determining a developed bias of an existing plurality of first user reviews for a first item. The method includes determining a tendency value of a designated user indicative of a tendency of a user sentiment exhibited in user reviews for respective second items provided by the designated user deviating from an average sentiment of the respective second items. The method includes determining an influential prompt in which the designated user provides an input for the first item, the influential prompt being offset by an offset value based on the developed bias and the tendency value. The method includes prompting the designated user with the influential prompt and receiving the input from the designated user. The method includes updating the tendency value based on the input.

Abstract: A method and system of evaluating a user experience (UX) design are provided. A UX design is received. All objects that are identified to be part of a background of the input UI screen are removed to create a filtered input UI screen. The input UI screen is assigned to a cluster. A target UI screen of the input screen is determined and its background removed, to create a filtered target UI cluster. The target UI screen is assigned to a cluster. The filtered input UI screen is used as an input to a deep learning model to predict a target UI cluster. The predicted target UI cluster is compared to the filtered target UI cluster based on the clustering. Upon determining that the filtered target UI cluster is similar to the target UI screen, the UX design is classified as being successful.

Abstract: One embodiment provides a method for augmenting missing values in historical or market data for deals. The method comprises receiving information relating to a set of deals. For any service included in one or more deals of the set of deals but not included in one or more other deals of the set of deals, the method further comprises augmenting, for any or all of the one or more other deals that does not include the service, one or more missing values for the service with one or more recommended values based on a recommendation algorithm. The service may be at any service level of a hierarchy of services.

Abstract: One embodiment provides a method comprising maintaining a weather model based on predicted weather conditions for an air traffic control zone. A hash table comprising multiple hash entries is maintained. Each hash entry comprises a timestamped predicted weather condition for a cell in the zone. A flight plan request for a drone is received. The request comprises a planned flight path for the drone. For at least one cell on the planned flight path, same latitude or same longitude cells, whichever is most closely orthogonal to a direction of the planned flight path, are heuristically probed. Weather conditions for the at least one cell are estimated based on predicted weather conditions for the same latitude or same longitude cells. An executable flight plan is generated if the planned flight path is feasible based on the estimated weather conditions; otherwise, a report including an explanation of infeasibility is generated instead.

Abstract: Embodiments relate to systematic explanation of neural model behavior and effective deduction of its vulnerabilities. Input data is received for the neural model and applied to the model to generate output data. Accuracy of the output data is evaluated with respect to the neural model, and one or more neural model vulnerabilities are identified that correspond to the output data accuracy. An explanation of the output data and the identified one or more vulnerabilities is generated, wherein the explanation serves as an indicator of alignment of the input data with the output data.

Abstract: A method and system of automatically identifying topics of a conversation are provided. An electronic data package comprising a sequence of utterances between conversation entities is received by a computing device. Each utterance is classified to a corresponding social action. One or more utterances in the sequence are grouped into a segment based on a deep learning model. A similarity of topics between adjacent segments is determined. Upon determining that the similarity is above a predetermined threshold, the adjacent segments are grouped together. A transcript of the conversation including the grouping of the adjacent segments is stored in a memory.

Abstract: One embodiment provides a method for top-down pricing of an in-flight deal. The method comprises receiving a first set of information relating to the in-flight deal. The in-flight deal comprises multiple services. The method comprises, for each service of the in-flight deal, selecting a corresponding set of peer deals from historical and market data based on the first set of information, and mining costs for the service from the corresponding set of peer deals. The method further comprises determining a set of price points for the in-flight deal based on each cost mined.