Natural-Language Patent Searching
 

Identify Relevant References fast

The Problem:

Patent searching is fundamentally a natural-language problem.  A search query, using conventional search software, attempts to capture the essence of the invention in a series of search terms.  Conventional search software identifies references containing the search terms from databases, containing issued patents and scientific literature and retrieves thousands of potentially relevant documents that must be compared with a description of the invention to determine the level of relevance.  This process that can take weeks to complete, costing patentees thousands of dollars in legal fees and lost inventor R&D time.  

Reduce Search and analysis time by 2/3

Natural Language SeArching:

Natural language queries can be used to make this process completely automated.  Using natural language searching, the inventor uploads a invention disclosure, product data sheet, or manuscript.  The computer "reads" the uploaded material, performs a search of relevant data bases identifying relevant documents, and then compares the identified documents with the uploaded material, scoring each reference and ranking them by relevancy.  The inventor and patent attorney are returned a list of highly relevant reference that have a high likelihood of impacting the invention.  

The TEAM

Curtis Wadsworth, J.D., Ph.D. is a long-time Pittsburgh-based patent lawyer and Ph.D. biochemist who has worked throughout his career to apply technology to the difficult problem of patent searching.

Michael Shamos, J.D., Ph.D. is a Distinguished Career Professor in the School of Computer Science at Carnegie Mellon University and Director of the MSIT eBusiness Technology Program

 
 

JOIN Our Mailing list