Some feel that artificial intelligence (AI) is revolutionizing the way patents are prosecuted, while others disagree. When we hear this term used in relation to patent prosecution, we immediately think of prior art searching. However, AI is being employed in several other ways.
As Susan Krelitz, adjunct professor of Intellectual Asset Management at the Mitchell Hamlin School of Law said, "IP Law firms and departments will use AI if it makes their life easier, faster or less expensive. It is that simple." Let us discuss each of the following areas, with specific examples, where AI has proven to be easier, faster or less expensive:
- Prior art searching
- Invention submission evaluation
- Watching the competition and identifying new competitors
- Directing R&D with white space landscaping
- Opposition Invalidity Searching
Prior art searching
The most obvious use of AI within the IP industry is prior art searching. AI and machine learning are technologies that allow us to easily and quickly gain insight into massive amounts of patent data. The traditional method of conducting a prior art search is to have a researcher construct a query into a patent database. The query will identify published patents that meet the specific criteria in the query. The criteria may include class codes, keywords, exclusionary words, etc.
The success of the prior art search is dependent on the skill of the searchers. Did they consider the right classes? Did they select the right keywords? Did they exclude keywords, without which the results will be overly inclusive? Did they use the correct Boolean operators? In other words, the search inquiry will return exactly what the researcher requested, nothing more and nothing less. The result of a traditional search is usually a long list of patents that the searcher must then sift through and prioritize.
AI-based searching is different. It processes human language with flexible semantics. There is no need to learn Boolean search structures. AI allows the user to input any description of an invention directly into the system, which then automatically extracts the meaning of the text and identifies patents with a similar purpose or technical content. The system intelligently analyses the data and is not dependent on the specific quires selected by the searcher. Take Octimine, for example. An innovative start-up in the field of IP management and recognized as one of the leading AI platforms, Octimine takes semantic patent search, analytics, and machine learning to the next level. Founded in 2015 by former Max-Planck-Institute and LMU Munich researchers, the company was acquired by the Dennemeyer Group in October 2018 and can assist its users, through AI, in various aspects of patent prosecution.
Octimine has a simple user interface but also uses a hybrid approach for more flexibility. The software solution allows a searcher to input natural language text in any format and refine the scope of the search by using specific filters. One example of a filter is a date range. Octimine users can restrict the search to prior art that existed on or before a particular date. Although the ease of entering a query is a definite advantage of AI, many see the way AI returns results as even more valuable. Traditional search engines return a long list of patents with little or no ranking and no visualization. Octimine ranks results by relevance and allows the searcher to quickly see the most relevant prior art.
Modern IP search tools also use visualization to illustrate search results and help users to get an overview of the results in a split second...click below to download and read the full article.
This article was first published in The Global IP Matrix magazine, issue 4, May 2019.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.