System and method for relation extraction using adaptive thresholding and localized context pooling (ATLOP). The system includes a computing device, the computing device has a processer and a storage device storing computer executable code. The computer executable code is configured to provide a document; embed entities in the document into embedding vectors; and predict relations between a pair of entities in the document using their embedding vectors. The relation prediction is performed based on an improved language model. Each relation has an adaptive threshold, and the relation between the pair of entities is determined to exist when a logit of the relation between the pair of entities is greater than a logit function of the corresponding adaptive threshold.

In some embodiments, the present disclosure provides for an exemplary computer-implemented system that may include a longitudinal data engine, including: a processor and specialized index generation software to generate: an index data structure for a respective event type associated with each respective subject or object; where each respective index data structure is a respective event type-specific data schema, defining how to store events of a particular event type to form longitudinal data of each respective subject or object; an ontology data structure that is configured to describe one or more properties of a respective event of a respective subject or object; and longitudinal data extraction software to extract a respective longitudinal data for a plurality of index data structures and a plurality of ontology data structures associated with a plurality of subjects or objects.

A system for providing a search service application is disclosed and includes an application builder component that provides a search model for a first object of a plurality of objects. The search model is based at least on an end-user input field corresponding to a first attribute of the first object and a search result output field corresponding to a second attribute of the first object. The search model is also associated with a backend data store that supports a storage structure that stores information relating to the first object. The system also includes a deployment engine that automatically configures a search engine system associated with the backend data store to place a portion of indexed data into a first partition and to place another portion of indexed data into at least another partition based on the search model.

One example method includes collecting annotated unstructured documents that each include a table with words whose respective column indices are known, using the documents to train a model to detect a table header in a given document, identifying, by the model, a region of a document that corresponds to a table header in a new document that is not part of the training data, using an algorithm to perform a segmentation process on the table header that identifies column boundaries in the table header, and to use the identified column boundaries to preliminarily assign a respective column index to each word in the table header. Finally, a graph neural network model is run on a graph that includes the words in the table, and running the graph neural network generates a refined prediction of a respective column index for each of the words in the table of the new document.

The disclosure provides a text classification model training method, a text classification method, an apparatus, an electronic device, and a computer-readable storage medium, and relates to artificial intelligence technology. The text classification model training method includes: performing machine translation on a plurality of first text samples in a first language to obtain a plurality of second text samples in a second language different from the first language; training a first text classification model for the second language based on a plurality of third text samples in the second language and corresponding class labels; performing confidence-based filtering on the plurality of second text samples by the trained first text classification model; and training a second text classification model for the second language based on the filtered second text samples.

A computer-implemented method, apparatus and program product use the spatial locations of words identified in an unstructured document to both reconstruct lines in the unstructured document and vertically partition the unstructured document. Key/value pairs may then be generated from one or more of the reconstructed lines by using one or more words to one side of the vertical partition as keys and using one or more words to the other side of the vertical partition as values.

Methods, systems and computer-readable media enable various techniques related to semantic navigation. One aspect is a technique for displaying semantically derived facets in the search engine interface. Each of the facets comprises faceted search results. Each of the faceted search results is displayed in association with user interface elements for including or excluding the faceted search result as additional search terms to subsequently refine the search query. Another aspect automatically infers new metadata from the content and from existing metadata and then automatically annotates the content with the new metadata to improve recall and navigation. Another aspect identifies semantic annotations by determining semantic connections between the semantic annotations and then dynamically generating a topic page based on the semantic connections.

The invention discloses a method and a device for sorting Chinese characters, searching for Chinese characters and constructing a dictionary, and relates to the technical field of computers. A specific implementation of the method includes: obtaining the first basic character-forming component of a Chinese character according to the stroke order as the First Character, and encoding the First Character to obtain the First Character code, where the First Character includes the first character-forming component and the first main stroke component of a Chinese character; obtaining the number of strokes included in each Chinese character, and obtaining the corresponding stroke string of each Chinese character; using the First Character code as the first and highest priority sorting field, the number of strokes as the second sorting field, and the stroke string as the third and the lowest priority sorting field to sort Chinese characters. This embodiment can solve the problem of difficulty in sorting and searching of Chinese characters caused by the unfixed definition and position of radicals.

A method is provided for an application program interface (API) to interface with an ontology store storing a plurality of modifiable ontology models having associated dynamic definitions associated that define classes of the associated ontology model and relationships between the respective classes and that is modifiable over time. The method includes receiving from a requesting entity a request that specifies an ontology model and one or more parameters defining attributes of an instantiated ontology object, accessing the ontology store, identifying an ontology model in the ontology store that corresponds to the ontology model specified, and manipulating the identified ontology model based on its one or more parameters. The method further includes generating a semantics query for accessing the identified ontology model based on the one or more parameters specified in the request, submitting the semantics query to and receiving query results from the ontology store, and returning the query results to the requesting entity.

The disclosure herein describes providing responses to natural language queries associated with transcripts at least by searching multiple indexes. A transcript associated with a communication among a plurality of speakers is obtained, wherein sets of artifact sections are identified in the transcript. A set of section indexes is generated from the transcript based on artifact type definitions. A natural language query associated with the transcript is analyzed using a natural language model and query metadata of the analyzed natural language query is obtained. At least one section index of the set of section indexes is selected based on the obtained query metadata and that selected at least one section index is searched. A response to the natural language query is provided including result data from the searched at least one search index, wherein the result data includes a reference to an artifact section referenced by the searched section index(es).