Sarthak Chauhan

Title: Sarthak Chauhan: Inventor in Long-Range Event Relation Extraction

Introduction

Sarthak Chauhan is an inventor based in Dehradun, India, whose patent work concerns artificial intelligence and natural language processing. His disclosed invention addresses how systems can identify relationships between events that appear far apart in a document.

Latest Patents

Chauhan is associated with 1 patent, titled “Generating and Utilizing Models for Long-Range Event Relation Extraction.” The invention describes creating a dataset by adding synthetic sentences to a digital document containing an event pair. A generative language model produces sentences that can be inserted between the events to meet a long-range relation threshold. The augmented dataset can then be used to train an event relation extraction model to identify relationships between distant events.

Career Highlights

Chauhan is associated with Adobe, Inc. His patent work reflects the application of generative language models and machine learning to document analysis, particularly the extraction of event relationships across longer passages.

Collaborations

The disclosed invention combines synthetic data generation with model training for long-range event relation extraction. The available information identifies Adobe, Inc. as Chauhan’s company but does not specify additional collaborators.

Conclusion

Sarthak Chauhan’s patent focuses on improving event relation extraction by using generated sentences and an augmented training dataset to help models connect events across a document.