Abstract: A system, method and computer program product for use in managing collaboration on documents having digital ink on a network of computing devices is disclosed. Each computing device has a processor and at least one system application for processing handwriting input under control of the processor. The system application displays, on a display associated with one of the computing devices, a document having digital ink based on a journal of the document, defines the journal to have journal entries associated with at least handwriting input to the document represented by the digital ink, and communicates the journal entries of the journal with one or more of the other networked computing devices displaying the document. The handwriting input associated with the journal entries is handwriting input to the document via the input interface of any of the computing devices displaying the document based on the communicated journal entries.

Abstract: A bipartite application implements a table auto-completion (TAC) algorithm on the client side and the server side. A client module runs a local model of the TAC algorithm on a user device and a server module runs a global model of the TAC algorithm on a server machine. The local model is continuously adapted through on-the-fly training, with as few as a negative example, to perform TAC on the client side, one document at a time. Knowledge thus learned by the local model is used to improve the global model on the server side. The global model can be utilized to automatically and intelligently extract table information from a large number of documents with significantly improved accuracy, requiring minimal human intervention even on complex tables.

Abstract: Attention weights in a hierarchical attention network indicate the relative importance of portions of a conversation between an individual at one terminal and a computer or a human agent at another terminal. Weighting the portions of the conversation after converting the conversation to a standard text format allows for a computer to graphically highlight, by color, font, or other indicator visible on a graphical user interface, which portions of a conversation led to an escalation of the interaction from an intelligent virtual assistant to a human customer service agent.

Abstract: A method includes receiving a request from a user device accessing a webpage, the request including a webpage uniform resource locator (URL) and a user device identifier. The method includes retrieving a list of events associated with the user device based on the device identifier. The method further includes retrieving sets of rules. Each set of rules indicates events and URLs that satisfy the set of rules. Each set of rules is associated with a template that includes link rendering data for rendering a link on the user device. The method includes identifying a set of rules that is satisfied by the received URL and events, transmitting link rendering data associated with the identified set of rules to the user device, and transmitting link routing data to the user device. The link routing data is configured to route the user device to an application state corresponding to the webpage.

Abstract: A detection is made that a user of a platform has provided first question text associated with a question via a user interface (UI). A first answer field format including a first data type associated with one or more first key terms of the question text is identified. The UI is updated to include first answer field options related to the first question text based on at least the first data type of the identified answer field format. In response to a detection that the user has provided second question text associated with the question via the UI, a second answer field format including a second data type associated with second key terms and at least one of the first key terms is identified. The UI is updated to include second answer field options related to the first question text and the second question text.

Abstract: A system and method for creating a dynamic electronic form are described. The system may include one or more processors that cause the system to perform create an electronic form with one or more data entry fields. The system may also obtain access to a plurality of datasets, where each dataset may include multiple entry fields and integrate at least one identified dataset with the electronic form. The system may further suggest at least one data input in the data entry field based on information input in the data entry field by a user. The data entry input suggested may be sourced from the identified dataset integrated to the electronic form.

Abstract: An approach is provided that receives a document and a document type of the document. The document type identifies a document category to which the received document belongs. A set of linguistic metrics are retrieved that correspond to the document type. A quality of the received document is automatically determined based on a set of linguistic features found in the document as compared to the retrieved set of linguistic metrics. The document is then ingested into a corpus that is utilized by a question-answering (QA) system. The ingestion of the document is based on the determined quality.

Abstract: In one embodiment, a deep fusion reasoning engine receives network telemetry data collected from a network. The deep fusion reasoning engine learns resource utilizations for different heuristic packages that can be used in the network to evaluate operation of the network. The deep fusion reasoning engine selects one of the heuristic packages based on the resource utilizations learned for the different heuristic packages. The selected heuristic package comprises a subservice and a set of rules to be evaluated. The deep fusion reasoning engine deploys the selected heuristic package for execution by a device in the network to evaluate operation of the network using the set of rules.

Abstract: A reactive spreadsheet in which the functionality of the reactive spreadsheet is decoupled from the visualization, where the same object or variable may appear in different tabs and panels while still retaining its identity.

Abstract: Media, methods, and systems are disclosed for automatically calculating predicted values using deep learning models. One or more input forms having a plurality of input form field values are received. The input form field values are automatically parsed into a set of computer-generated candidate standard field values. The set of candidate standard field values are automatically normalized into a data frame having a set of normalized field values. In response to determining which portions of the data frame to apply to a line calculation neural network, a corresponding line calculation neural network is applied. At least one output form line calculation is performed. A natural language explanation regarding the at least one output form line calculation is generated. Additional user provided inputs are received in response to the natural language explanation, and at least one remaining calculation is carried out based on the one or more additional user provided inputs.

Abstract: Systems and methods may be provided for performing a search on an input text block. The input text block may be split into a plurality of input text segments. A text similarity algorithm may be used to find similar stored text segments to each of the plurality of input text segments.

Abstract: In an approach to interpreting and summarizing online terms of use, a computer receives a terms of use agreement from a data source and a request for interpretation of the terms of use agreement from a user. A computer categorizes the terms of use agreement into one or more categories based on a type of data source of the terms of use agreement. A computer ranks one or more words in the terms of use agreement based on the categorization and on one or more additional terms of use agreements of the type of data source. A computer generates a summary of the terms of use agreement based on the ranking. A computer displays the summary of the terms of use agreement to the user. A computer receives input from the user associated with consent to the terms of use agreement. A computer stores the input.

Abstract: Described herein are systems and methods that search videos and other media content to identify items, objects, faces, or other entities within the media content. Detectors identify objects within media content by, for instance, detecting a predetermined set of visual features corresponding to the objects. Detectors configured to identify an object can be trained using a machine learned model (e.g., a convolutional neural network) as applied to a set of example media content items that include the object. The systems provide user interfaces that allow users to review search results, pinpoint relevant portions of media content items where the identified objects are determined to be present, review detector performance and retrain detectors, providing search result feedback, and/or reviewing video monitoring results and analytics.

Abstract: A method and system for providing replacement text segment suggestions for an unoriginal text segment in a document may include examining a portion of the document to determine if the portion includes a text segment containing unoriginal content. Upon determining that the portion includes the unoriginal text segment, the method may enable display of a notification that the text segment contains unoriginal content and receive a request, via a network, to provide the replacement text segment for the unoriginal text segment. Upon receiving the request, the method may identify the replacement text segment for the text unoriginal segment, based at least in part a guideline relating to the use of content that is included in a source. The replacement text segment may include a citation for the source.

Abstract: In an approach, a processor trains a statistical classifier and a set of micro classifiers. A processor receives an input to be classified by the statistical classifier. A processor receives a label assigned to the input by the statistical classifier and respective labels assigned by each micro classifier of the set of micro classifiers. A processor determines that the label assigned by the statistical classifier is the same as at least one label assigned by at least one micro classifier of the set of micro classifiers. A processor generates a natural language explanation for assigning the label using the at least one micro classifier and the label. A processor outputs the label and the natural language explanation to a user of a computing device. A processor receives user feedback from the user in the form of an acceptance or a rejection of the natural language explanation.

Abstract: Embodiments disclose a method for automatic summarization of content. The method includes accessing a plurality of stories from a plurality of data sources for a predefined time. Each story is associated with a media item. The method includes plotting the plurality of stories over the predefined time for determining one or more peaks and extracting a set of stories from the one or more peaks. The method includes detecting one or more themes from the set of stories using LDA algorithm. Each theme is associated with a group of stories. The method further includes determining at least one subset of stories for each theme from the group of stories representing the set of stories in the one or more peaks using RBM algorithm. The method includes generating a summarized content for each user based on an associated user profile and the at least one subset of stories.

Abstract: The present disclosure relates to systems and methods for analyzing citationally related content and identifying, based on the analysis, a risk of impliedly overruled content. Embodiments provide for receiving case law data from a document source, for extracting a case triple that includes a first case overruling or abrogating a second case, and a third case citationally related to the second case. Features may be generated from case triple, such as natural processing language features comparing the language in the various cases of the triple, and feeding the generated features to a main classifier. In embodiments, the main classifier classifies the case triple into a class indicating the risk probability that the second case is impliedly overruled by the first case.

Abstract: Methods and systems for structuring, storing and displaying time series data in a user interface. One system includes processors executing instructions to determine, from time series data from a first sensor, a first subset of time series data for the first batch from the first start time and the first end time, determine, from the time series data from the first sensor, a second subset of time series data for the second batch from the second start time and the second end time, generate a time series user interface comprising a chart, the chart including a first plot for the first subset of time series data and a second plot for the second subset of time series data, the first plot being aligned to the second plot, and cause presentation of the time series user interface.

Abstract: Systems for generating a USPTO office action response document based on an input USPTO Office Action document. A method implementation for constructing the USPTO office action response traverses a ruletree that has been dynamically formed from user answers to successive questions based on characteristics of the USPTO Office Action document. The ruletree comprises a first set of data items that pertains to an HCI and the ruletree further comprises a second set of data items that pertain to the target USPTO office action response document. Correspondence between the set of user answers to questions and passages in the USPTO office action response document are maintained in the ruletree. The ruletree expands and contracts during user interaction. To generate the USPTO office action response document, the ruletree is traversed in a sorted order, where traversal of a node generates a passage or passages in the generated USPTO office action response document.

Abstract: An intelligent text cleaning method includes: acquiring a text set, and preprocessing the text set to obtain a word vector text set; subjecting the word vector text set to a full-text matrix numeralization to generate a principal word vector matrix and a text word vector matrix; inputting the principal word vector matrix to a BiLSTM model to generate an intermediate text vector; inputting the text word vector matrix to a convolution neural network model to generate a target text vector; and concatenating the intermediate text vector and the target text vector to obtain combined text vectors, inputting the combined text vectors to a pre-constructed semantic recognition classifier model, outputting an aggregated text vector, subjecting the aggregated text vector to reverse recovery using a word2vec reverse algorithm, and outputting a standard text. The present application realizes accurate text cleaning.