On April 18, the Federal Circuit Court released its  decision in Recentive Analytics, Inc., v. Fox Corp., Fox Broadcasting Company, LLC, and Fox Sports Productions, LLC. In Recentive, the patents claimed the use of machine learning to solve problems confronting broadcasters, such as the Defendants (collectively “Fox”), related to determining the scheduling of live events and optimizing network maps, which determine the programs or content displayed by a broadcaster’s channels within certain geographic markets at particular times. The Federal Circuit held that certain types of machine learning patents are per se patent-ineligible. However, the Court also noted that the decision does not apply to all machine-learning patents.

This decision arose from an appeal of a District Court decision on a motion to dismiss by Fox. In that decision, the District Court ruled that the four patents asserted by Recentive Analytics were not direct to patent-eligible subject matter. The Federal Circuit affirmed the decision of the District Court, stating that “today, we hold only that patents that do no more than claim the application of generic machine learning to new data environments, without disclosing improvements to the machine learning models to be applied, are patent ineligible under § 101.”

Many have questioned whether the holding in Recentive heralds the end of patenting machine learning related inventions. However, Recentive is consistent with previous decisions by the Federal Circuit on subject-matter eligibility. As an example, independent claim 1 of U.S. Patent No. 10,911,811 (one of the four patents asserted) recites:

A computer-implemented method for dynamically generating a network map, the method comprising:
receiving a schedule for a first plurality of live events scheduled to start at a first time and a second plurality of live events scheduled to start at a second time;
generating, based on the schedule, a network map mapping the first plurality of live events and the second plurality of live events to a plurality of television stations for a plurality of cities,
wherein each station from the plurality of stations corresponds to a respective city from the plurality of cities,
wherein the network map identifies for each station (i) a first live event from the first plurality of live events that will be displayed at the first time and (ii) a second live event from the second plurality of live events that will be displayed at the second time, and
wherein generating the network map comprises using a machine learning technique to optimize an overall television rating across the first plurality of live events and the second plurality of live events;
automatically updating the network map on demand and in real time based on a change to at least one of (i) the schedule and (ii) underlying criteria,
wherein updating the network map comprises updating the mapping of the first plurality of live events and the second plurality of live events to the plurality of television stations; and
using the network map to determine for each station (i) the first live event from the first plurality of live events that will be displayed at the first time and (ii) the second live event from the second plurality of live events that will be displayed at the second time.

In this claim, if one were to replace the bolded phrase “machine learning technique” with “human,” the entire method could be implemented by a person using pen-and-paper. The computer generally, and machine learning specifically, is only recited as the mechanism or vehicle by which the abstract idea is applied or implemented. Viewed another way, it could be said the claim anthropomorphizes the computer, which is often an indication that the claim may be too general to be patent eligible. It is therefore unsurprising that this claim was found to be subject-matter ineligible by the court, as well as claims with similar scope in the other asserted patents.

There are several conclusions for clients to consider moving forward.

First, inventions that use or rely on machine learning are still patentable. The Federal Circuit limited their holding only to patents that do no more than claim the application of generic machine learning to new data environments, without disclosing improvements to the machine learning models to be applied. So, the door is still open to machine learning patent applications that disclose, and claim, improvements to the machine learning models being used. Moreover, other tests for subject-matter eligibility, such as the machine-or-transformation test, are still good law and therefore offer alternative avenues for proving patent-eligibility. 

Second, patents that simply claim the application of a business method or other abstract idea using a machine learning model are unlikely to be granted or to survive motions to dismiss in litigation. This is consistent with prior case law on subject matter eligibility. For example, in Alice Corp. v. CLS Bank Int’l, 573 U.S. 208 (2014), the Supreme Court stated that it is not enough to render a claim patent-eligible by stating an abstract idea and adding the words “apply it with a computer.” The Federal Circuit’s decision in Recentive is consistent with the holding in Alice – holding the claims that recite an abstract idea applied with a machine learning model to similarly be patent ineligible. 

How significant these consequences from Recentive prove to be will remain to be seen and will probably be resolved through litigation over the next few years. We predict that machine learning patent applications will require a great degree of specificity in the claim language and the detailed description in order to be granted.

Thomas | Horstemeyer has robust experience in patent writing and patent prosecution. For questions and guidance, email us at info@thip.law.