This inventor holds 1 USPTO granted patent and 1 published patent application. Top assignee: Microsoft Technology Licensing, LLC. Active years: 2025.
Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile
Company Filing History:
Years Active: 2025
Title: Christopher H. Lin: Inventor in Semantic Parsing
Introduction
Christopher H. Lin is an inventor based in Bellevue, Washington. His work includes a patent on using semantic parsing to turn natural-language utterances into executable programs. Lin is associated with Microsoft Technology Licensing, LLC, and has 1 patent listed in the information provided.
Latest Patents
Lin’s patent, “Semantic Parsing of Utterance Using Contractive Paraphrasing,” describes systems and methods for generating a program from a natural-language utterance. The approach translates the utterance into logical-form instructions that an application can execute.
The method uses a pretrained language model to create a canonical utterance as an intermediate step before producing the logical form. It can prompt the model with example input-and-output pairs and apply constrained decoding, checking predicted word sequences against rules for canonical utterances. The resulting canonical utterance is then used to generate a program for execution.
Career Highlights
The patent reflects work at the intersection of natural-language processing, semantic parsing, and program generation. Its described techniques use a transformer-based, autoregressive language model to paraphrase utterances and support the conversion of natural language into executable instructions.
Collaborations
The information provided names Samuel McIntire Thomson and Emmanouil Antonios Platanios as Lin’s coworkers. Their inclusion here does not establish that they were co-inventors or collaborators on the patent.
Conclusion
Christopher H. Lin’s patent describes a language-model-based approach to converting natural-language utterances into programs through canonical paraphrasing and semantic parsing.
Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile
