Comparing Complex Concepts with Transformers: Matching Patent Claims Against Natural Language Text

An entity defending itself against infringement may attempt to A key capability in managing patent applications or a patent invalidate a patent by finding novelty-destroying prior art to that portfolio is comparing claims to other text, e.g. a patent patent. In all cases, the key task is to search through a set of specification. Because the language of claims is different from documents and determine whether those documents cover all language used elsewhere in the patent application or in non-patent aspects of each claim of the subject patent application or granted text, this has been challenging for computer based natural patent. Thus, a claim of a subject patent (application) may be language processing. We test two new LLM-based approaches considered a query to an information retrieval system whose and find that both provide substantially better performance than objective is to retrieve a document or set of documents that previously published values. The ability to match dense contain all aspects of that claim.