A single critical prior art missed by a searcher may come to haunt the patent applicant/patentee later as it may hamper the patent prosecution or become a ground for invalidation or infringement. Therefore, it is important to ensure that not a single relevant term is omitted as it would result in missing important prior art. Hence, it has become imperative to integrate artificial intelligence-driven semantic search with conventional keyword searches.
Patents are difficult to read, and many are written in a manner that is vague and tough to understand and search. Every so often, drafters employ novel terminology when drafting a patent application, which is not readily known at the time of the search. Consequently, a conventional, patent searching approach involving keywords and phrases grouped in a query using Boolean and proximity operators may not display a critical prior art in the search result.
Semantic search works over the contextual meaning of the input sentence or paragraph and consequently can help uncover such missed results. Furthermore, you can easily NOT the results of the semantic search with your Boolean query to see the delta results and scan them for any relevant record.
For prior art or invalidity searches with a short turnaround time, you can also directly start with a semantic search. In many cases, you will discover the matching set of the prior art with just this approach. Both these types of searches might not need 100% precision, and if you can plead your case with a portion of the relevant prior art, then further searches become redundant.
Click here to read the original article published by Gridlogics.
Please give your feedback on this article or share a similar story for publishing by clicking here.
