Sign up for our newsletter and get the latest big data news and analysis.

The Deep Learning Patent Land Rush

The number of “deep learning” patents is exploding. Since I’m the sole or lead co-inventor on eight of these patents, I’ve been keeping track. In fact, I have a list of such patents (in numerical order) which ranges from January 1, 2013 to December 3, 2019, and which, on a year-by-year, results in an eye-popping data visualization below:

More specifically, the Year-over-Year results are as follows: 2013 (3 patents), 2014 (4), 2015 (4), 2016 (36), 2017 (83), 2018 (162), and in 2019 through December 3, 2019 (361). The number of deep learning patents has been doubling every year for the last five years.

The Leader Board validates my search terms—where the terms “deep learning” or “deep neural” or “multi-layer neural” can be found in a patent’s Claims—because “the usual suspects” are dominate the list. As of December 3, 2019, the Leader Board is:

  1. IBM (51)
  2. Google (39)
  3. Microsoft (28)
  4. Siemens Healthcare (27)
  5. Samsung (16)
  6. NEC (14)
  7. Amazon (13)
  8. Baidu (12)
  9. Facebook (12)
  10. Ford (11)
  11. Intel (11)
  12. Adobe Systems (10)

That said, my understanding prior to June 27, 2016, was that, due to the U.S. Supreme Court decision in Alice Corp. v. CLS Bank International [1], the United States Patent and Trademark Office (USPTO) was going to reject every patent application for a “software system” on the grounds that a system was a set of tasks that a computer could do, and so was patent ineligible.

But on June 27, 2016, the Federal Circuit Court of Appeals published its decision in Bascom Global v. AT&T Mobility.[2] I was lucky to see it. In his Bascom Global opinion, Judge Raymond Chen explained that the patent at issue had passed a “step two” test in Alice: “As is the case here,” he wrote, “an inventive concept can be found in the non-conventional and non-generic arrangement of known, conventional pieces.” [3]

Because U.S. patent laws now favor the “first to file,” I was energized write a provisional application for “Using Classified Text and Deep Learning Algorithms to Identify Risk and Provide Early Warning.” My provisional was filed on July 1, 2016. That’s the Priority Date for each of the eight patents in my “family.”

After July 1, 2016, the next set of events were unanticipated and surprising.

On September 27, 2016, my formal application was filed;

On December 12, 2016, the USPTO issued its first Office Action, which was a Notice of Allowance (!); and

On January 24, 2017, my first patent issued as U.S. Pat. No. 9,552,548.

January 24, 2017 is three days short of four months after the formal application was filed on September 27, 2016. Light speed for the USPTO! Most inventors wait years for a patent to go through the process, not months, so what explains this? The explanation is that I had a patent superpower: The Petition to Make Special.

Here’s one way to qualify for a Petition to Make Special: “An application may be made special upon filing … a statement from a registered practitioner that he or she has evidence that the applicant is 65 years of age or older …” [4] That was stunning. I was already “65 years of age, or more.” Better still is that this superpower works even when a co-inventor is not 65 years of age or older. I’ve had two patents processed to completion where each of my co-inventors was in their thirties or forties. In each instance, my Petition was granted.

The three takeaways are: (1) deep learning patents are booming; (2) software system patents are possible; and (3) there’s a wonderful advantage to having a knowledgeable senior on your innovation team.


[1] Alice Corp. Pty. Ltd. v. CLS Bank International, 573 U.S. 208, 134 S. Ct. 2347 (2014).

[2] Bascom Global Internet Servs., Inc. v. AT & T Mobility, LLC, 827 F.3d 1341 (Fed. Cir. 2016).

[3] Ibid. at 1350.

[4] 37 CFR 1.102(c) and MPEP § 708.02.

About the Author

Nelson E. (Nick) Brestoff holds a B.S. degree in Engineering Systems from UCLA, an M.S. degree in Environmental Engineering Science from the California Institute of Technology, and a J.D. degree from the USC Gould School of Law. He was a California litigator for 38 years. He is the sole or lead co-inventor on eight deep learning patents, now assigned to Intraspexion LLC, and is author of AI Concepts for Business Applications.

Sign up for the free insideBIGDATA newsletter.

Leave a Comment

*

Resource Links: