International Business Machines Corporation

Riddhiman Dasgupta

Sameep Mehta

Arun Kumar

Berthold Reinwald

Arvind Agarwal

Hima Patel

Vitobha Munigala

Balaji Ganesan

Akshay Parekh

Aspects of the present disclosure relate to neural-based ontology generation and refinement. A set of input data can be received. A set of entities can be extracted from the set of input data using a named-entity recognition (NER) process, each entity having a corresponding label, the corresponding labels making up a label set. The label set can be compared to concepts in a set of reference ontologies. Labels that match to concepts in the set of reference ontologies can be selected as a candidate concept set. Relations associated with the candidate concepts within the set of reference ontologies can be identified as a candidate relation set. An ontology can then be generated using the candidate concept set and candidate relation set.

Neural-based ontology generation and refinement

One embodiment provides a method, including: obtaining at least two documents, wherein one of the at least two documents comprises a different revision of another of the at least two documents; identifying a structure of each of the at least two documents by parsing each of the at least two documents to extract text from each of the at least two documents; aligning sections of the at least two documents, wherein the aligning comprises matching a section from one of the at least two documents and a corresponding section from another of the at least two documents; identifying at least one difference between the at least two documents; assigning a semantic label to the identified at least one difference; and providing a summary of the identified at least one difference by compressing the text surrounding the identified at least one difference using the assigned semantic label.

Document revision change summarization

Growing community of inventors

Kolkata, India

Riddhiman Dasgupta