District Court Decision Provides Guidance on Patent Eligibility of Claims Directed to Using and Training Machine Learning Models | Akin Gump Strauss Hauer & Feld LLP

Judge Williams within the District of Delaware just lately granted a movement to dismiss the criticism as a result of the patents-in-suit declare ineligible material underneath 35 U.S.C. § 101. The patents are directed to utilizing and coaching machine studying fashions for producing community maps (or tv schedules) and optimizing occasion schedules. The court docket discovered the claims invalid as a result of they recite the summary concepts of producing community maps and occasion schedules utilizing generic mathematical methods.

Recentive Analytics, Inc. v. Fox Corp., No. 22-1545-GBW (D. Del.).

Plaintiff Recentive Analytics sued Fox Corp. for infringing U.S. Patent Nos. 10,911,811 and 10,958,957 (the “Network Map Patents”) and U.S. Patent Nos. 11,386,367 and 11,537,960 (the “Machine Learning Training Patents”).

The Network Map Patents are directed to a computer-implemented methodology of receiving a schedule of occasions in numerous time slots, assigning these occasions for every slot to a number of TV stations, utilizing machine studying to optimize TV rankings, and updating the community map on demand and in actual time. The consultant declare of the Network Map Patents recites 4 steps: (1) a accumulating step, i.e., receiving schedules of occasions; (2) an analyzing step, i.e., utilizing a machine studying algorithm to create a community map; (3) an updating step, i.e., updating the community map based mostly on real-time data; and (4) a utilizing step, i.e., utilizing the community map to decide for every station which occasion shall be proven.

The Machine Learning Training Patents are directed to a computer-implemented methodology of producing an occasion schedule by way of a machine studying mannequin that has been iteratively skilled to optimize goal options (e.g., attendance, revenue, income, bills) based mostly on enter parameters (e.g., venue availability, venue areas, ticket costs, performer charges, venue charges, scheduled performances). The consultant declare of the Machine Learning Training Patents additionally recites 4 steps: (1) a accumulating step, i.e., receiving occasion parameters and goal options; (2) a coaching step, i.e., feeding the info right into a machine studying mannequin and coaching it to determine relationships; (3) an output step, i.e., inputting traits of future reside occasions and receiving from the machine studying mannequin an optimized schedule; and (4) an updating step, i.e., detecting adjustments to the inputs and feeding these inputs to the machine studying mannequin to re-optimize the schedule.

The court docket analyzed eligibility utilizing the Supreme Court’s two-step Alice framework. In the 1st step, a court docket determines whether or not the claims are “directed to a patent-ineligible idea,” equivalent to an summary thought. Alice Corp. Pty. Ltd. v. CLS Bank Int’l, 573 U.S. 208, 217 (2014). If they’re, the court docket proceeds to step two and considers “the weather of every declare each individually and ‘as an ordered mixture’ to decide whether or not the extra components ‘rework the character of the declare’ right into a patent-eligible software.” Id. (quoting Mayo Collaborative Servs. et al. v. Prometheus Laby’s., Inc., 566 U.S. 66, 78-79 (2012)).

Alice Step One

Addressing Alice the 1st step, the court docket discovered that the claims are directed to a patent-ineligible summary concepts of producing community maps and occasion schedules utilizing identified generic mathematical methods. The court docket in contrast the claims to the these in Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350 (Fed. Cir. 2016), as a result of they’re directed to accumulating data, analyzing it and displaying sure outcomes of the gathering and evaluation. Recentive made three arguments to differentiate the patent claims from these beforehand discovered to be ineligible, every of which the court docket rejected.

First, Recentive argued that machine studying algorithms course of data in another way from the human mind, in that people course of information qualitatively fairly than quantitatively. The court docket discovered, nevertheless, that it’s irrelevant whether or not a human making a community map would run a help vector machine of their mind. Instead, the court docket determined that the related query is whether or not the machine studying processes are algorithms, and as a result of machine studying is algorithmic in nature, the patents are directed to an summary thought.

Second, citing SRI International, Inc. v. Cisco Systems, Inc., 930 F.3d 1295 (Fed. Cir. 2019), Recentive argued that the claims are patent eligible as a result of the claimed processes require an excessive amount of information and computing energy for the human mind to do. In SRI, the Federal Circuit discovered claims eligible as a result of the human thoughts was not geared up to interact in community monitoring of particular community packets. The court docket distinguished SRI as a result of people can interact in mathematical methods to carry out machine studying, albeit slowly. The court docket additionally discovered that the patents-in-suit don’t enhance technical functioning. Rather, they merely use a pc as a instrument to carry out community mapping and occasion scheduling. The court docket additionally relied on Trinity Info Media, LLC v. Covalent, Inc., 72 F.4th 1355 (Fed. Cir. 2023), which held {that a} human being incapable of matching processing velocity doesn’t make an summary course of patent eligible.

Third, Recentive analogized the patents-in-suit to these in McRO, Inc. v. Bandai Namco Games America Inc., 837 F.3d 1299 (Fed. Cir. 2016), which held that the use of an unconventional rule set distinguished the patents from prior artwork human strategies if the applying of the foundations created a tangible consequence (the sequence of animated characters in McRO). The court docket disagreed that McRO is analogous as a result of of the requirement in McRO that the foundations be “unconventional” and as a result of the Federal Circuit has been hesitant to increase McRO past its details.

The court docket discovered this case extra analogous to In re Board of Trustees of Leland Stanford Junior University, 991 F.3d 1245 (Fed. Cir. 2021) (“Stanford”) and SAP America, Inc. v. InvestPic, LLC, 898 F.3d 1161 (Fed. Cir. 2018), during which the Federal Circuit discovered claims to be patent ineligible. In Stanford, the claims recited a computerized methodology of inferring sure genetic information throughout sequencing, and the Federal Circuit distinguished McRO as a result of it concerned “sensible, technological enhancements extending past enhancing the accuracy of a mathematically calculated statistical prediction.” In SAP, the claims recited statistically analyzing funding data and reporting the outcomes, and the Federal Circuit distinguished McRO as a result of it was directed to “the creation of one thing bodily,” in contrast to the quantitative predictions in SAP.

Applying Stanford and SAP, the court docket discovered that the community maps and schedules are extra analogous to the tangibility degree in SAP’s monetary fashions than the animated characters in McRO. The court docket additionally discovered that altering a subjective course of the place artists are attempting to make a bit of artwork look good into an algorithmically pushed one targeted on quantitative prediction (as in McRO) is distinct from a course of the place each people and algorithms are attempting to maximize TV rankings. The court docket additionally famous that McRO claimed “particular and unconventional” guidelines, whereas the foundations within the patents-in-suit are admittedly typical machine studying methods described in broad purposeful phrases.

Addressing Alice step two, the court docket determined that the patents-in-suit don’t recite any creative idea as a result of the machine studying limitations are described solely in broad purposeful phrases and present little steerage on mannequin parameters or coaching approach. The Network Map Patents recite “utilizing a machine studying approach” within the claims and disclose utilizing “any appropriate machine studying approach.” The Machine Learning Training Patents recite and describe utilizing both a neural community or a help vector mannequin and iteratively coaching it. The court docket discovered that the patents additionally declare solely generic and typical computing gadgets.

Recentive argued that “the use of machine studying algorithms to generate community maps and optimize occasion schedules” is the creative idea. Recentive relied on Amdocs (Israel) Ltd. v. Openet Telecom, Inc., 841 F.3d 1288 (Fed. Cir. 2016), which held eligible patent claims relating to managing information over massive networks after they contained “particular enhancing limitations that essentially included the invention’s distributed structure.” The court docket discovered, nevertheless, that in contrast to the “unconventional technological answer (enhancing information in a distributed vogue)” in Amdocs, it’s undisputed that Recentive didn’t invent machine studying. In different phrases, the court docket concluded that the creative idea that Recentive identifies is merely the summary thought—making use of machine studying to optimization of community maps and occasion schedules.

Practice Tip: Patent Owners ought to keep away from describing and claiming the advance over the prior artwork in purely purposeful phrases in a result-oriented approach that quantities to encompassing the summary answer irrespective of how carried out. Instead, Patent Owners ought to describe and declare technical particulars for tangible elements within the claimed system, together with how the advance over the prior artwork is carried out. In the context of machine studying, Patent Owners ought to keep away from claims which are directed to generic machine studying methods for accumulating and analyzing data. Instead, Patent Owners ought to describe and declare particular and unconventional technological enhancements to machine studying techniques (e.g., within the mannequin parameters or coaching methods) that create a tangible consequence.


Recommended For You