In a first case of its kind, Health Discovery Corp. v. Intel Corp. on Dec. 27, the U.S. District Court for the Western District of Texas found claims of machine-learning patents invalid under Title 35 of the U.S. Code, Section 101, in a motion to dismiss filed under Federal Rule of Civil Procedure 12(b)(6).
This decision, on one hand, provides a road map that skilled counsel can follow to draft patents that are more likely to withstand eligibility challenges, but, on the other hand, could make patenting artificial intelligence inventions more nuanced absent due care.
The invention in the Health Discovery patents — U.S. Patent Nos. 7,117,188, 7,542,959, 8,095,483 and 10,402,685 — relates to using machine-learning systems to identify patterns in biological systems, such as genes, proteins, lipids, etc.
The machine-learning technology is a support vector machine, or SVM, and recursive feature elimination, or RFE. The claims are directed to training an SVM on training data and then using the trained SVM to find patterns in live data.
The court analyzed the claims using the two-step framework set forth in the U.S. Supreme Court's 2014 Alice Corp. v. CLS Bank International decision.
In Alice step one, the court determined that the claims were directed to an abstract mathematical concept of SVM-RFE. This is because the claims recited a mathematical technique that ranked and eliminated features using the SVM-RFE.
Excerpted from Law360. To read the full article, click here.