Anna Ronkainen, LL.M., of Onomatics, Inc., has posted Intelligent Trademark Analysis: Experiments in Large-Scale Evaluation of Real-World Legal AI, a paper she presented this week at ICAIL 2013: International Conference on Artificial Intelligence and Law.
Here is the abstract:
The intelligent trademark analysis system developed by Onomatics is a trademark information system based on an AI model of trademark similarity (likelihood of confusion). The basic technology can be used as a general trademark search engine as well as for more specific purposes ranging from trademark candidate ranking (TrademarkNow NameRank) to comprehensive risk analysis (TrademarkNow NameCheck) and IPR enforcement. This paper presents a brief overview of the system and the concrete applications available as of April 2013. The system has been evaluated on a large number of actual trademark opposition cases, over 30 000 from the USPTO TTAB and over 20 000 from the OHIM Opposition Division, currently with a precision of 79.9% and a recall of 94.9%.