Toru Yamazato

Patent Trials and Appeals Board

Sep 16, 2019

15136314 - (D) (P.T.A.B. Sep. 16, 2019)

UNITED STATES PATENT AND TRADEMARK OFFICE
UNITED STATES DEPARTMENT OF COMMERCE
United States Patent and Trademark Office
Address: COMMISSIONER FOR PATENTS
P.O. Box 1450
Alexandria, Virginia 22313-1450
www.uspto.gov
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR ATTORNEY DOCKET NO. CONFIRMATION NO.
15/136,314 04/22/2016 Toru YAMAZATO 20419.0046US01 8215
52835 7590 09/16/2019
HAMRE, SCHUMANN, MUELLER & LARSON, P.C.
45 South Seventh Street
Suite 2700
Minneapolis, MN 55402-1683
EXAMINER
CHANG, LI WU
ART UNIT PAPER NUMBER
2124
NOTIFICATION DATE DELIVERY MODE
09/16/2019 ELECTRONIC
Please find below and/or attached an Office communication concerning this application or proceeding.
The time period for reply, if any, is set in the attached communication.
Notice of the Office communication was sent electronically on above-indicated "Notification Date" to the
following e-mail address(es):
PTOMail@hsml.com
PTOL-90A (Rev. 04/07)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
____________________

Ex parte TORU YAMAZATO
____________________

Appeal 2018-008697
Application 15/136,314
Technology Center 2100
____________________

Before JENNIFER S. BISK, LARRY J. HUME, and
JULIET MITCHELL DIRBA, Administrative Patent Judges.

DIRBA, Administrative Patent Judge.

DECISION ON APPEAL1
Appellant2 seeks our review under 35 U.S.C. § 134(a) of the
Examiner’s rejection of claims 1, 4–6, 9–11, 14, 15, 18, and 19, which are
all claims pending in the application. Although Appellant appeals from a
non-final rejection, we have jurisdiction pursuant to 35 U.S.C. §§ 6 and 134

1 This Decision uses the following abbreviations: “Spec.” for the original
specification, filed April 22, 2016, claiming the benefit of an earlier foreign
patent application; “Non-Final Act.” for the Non-Final Office Action, mailed
August 30, 2017; “Br.” for Appellant’ Appeal Brief, filed April 23, 2018;
and “Ans.” for Examiner’s Answer, mailed June 20, 2018. Appellant
elected not to file a Reply Brief in response to the findings and conclusions
in the Examiner’s Answer.
2 According to Appellant, the real party in interest is the inventor, Toru
Yamazato. Br. 2.
Appeal 2018-008697
Application 15/136,314

2
because the claims have been twice presented and rejected. See Ex parte
Lemoine, 46 USPQ2d 1420, 1423 (BPAI 1994) (precedential).
We affirm.
BACKGROUND
Appellant’s disclosed embodiments and claimed invention relate to a
nonlinear optimization routine. Spec. 1–3, 55.
Claim 1, reproduced below, is illustrative of the claimed subject
matter:
1. A computer-implemented method for training an
artificial neural network with increased processing speed,
comprising:
[(1)] executing a search to improve a vector x so that the vector
x minimizes or maximizes a value of a function of x, f(x), the
vector x corresponding to an error function for training the
artificial neural network, the search including:
[(2)] assigning a vector g a gradient of f(x);
[(3)] executing a plurality of search steps, each search step
including:
[(4)] a) using the vector g to determine a search direction
vector, determining a step size using a parabolic
approximation, and updating the vector x based on the
determined search direction vector and the determined
step size, wherein determining the step size includes a
product of the search direction vector as determined and
a first derivative of the function f(x), and includes a
product of the search direction vector as determined and
a second derivative of the function f(x) approximated as a
parabolic function;
[(5)] b) using the vector g, the search direction vector as
determined in a), and the step size as determined in a) to
approximate a gradient of f(x), wherein the
approximation includes the product of the search
Appeal 2018-008697
Application 15/136,314

3
direction vector as determined in a) and the second
derivative of the function f(x) approximated as a
parabolic function, the approximation increasing the
processing speed for training the artificial neural network
by minimizing a number of calculations on derivatives of
f(x);
[(6)] c) determining if a convergence criterion is
satisfied; and
[(7)] d) if the convergence criterion is not satisfied,
assigning the gradient of f(x) as approximated in b) to the
vector g and returning to step a);
[(8)] if the convergence criterion is satisfied, using the vector x
as a value of a parameter of the artificial neural network.
Br. 23–24 (Claims App’x) (numbering added).
THE REJECTION
Claims 1, 4–6, 9–11, 14–15 and 18–19 stand rejected under 35 U.S.C.
§ 101 as directed to patent-ineligible subject matter.3 Non-Final Act. 2–6.
ANALYSIS
We review the appealed rejections for error based upon the issues
identified by Appellant and in light of the arguments and evidence produced
thereon. Ex parte Frye, 94 USPQ2d 1072, 1075 (BPAI 2010) (precedential).
To the extent Appellant has not advanced separate, substantive arguments

3 The header for the rejection identifies claims “9–1” (Final Act. 2), but the
rejection addresses claims 9, 10, and 11 (id. at 5). Appellant acknowledges
that independent claim 11 has been rejected (Br. 10, 18–21) and submits that
the dependent claims, such as claim 10, stand or fall with the respective
independent claims (id. at 10). As such, we consider the reference to claims
“9–1” to be a typographical error referencing claims 9–11.
Appeal 2018-008697
Application 15/136,314

4
for particular claims, or other issues, such arguments are waived. 37 C.F.R.
§ 41.37(c)(1)(iv).
We have considered all of Appellant’s arguments and any evidence
presented. We highlight and address specific findings and arguments for
emphasis in our analysis below.
Rejection of Claims 1, 4–6, 9–11, 14–15 and 18–19 under 35 U.S.C. §101
Section 101 of the Patent Act provides that “any new and useful
process, machine, manufacture, or composition of matter, or any new and
useful improvement thereof” is patent eligible. 35 U.S.C. § 101. But the
Supreme Court has long recognized implicit exceptions to this section:
“‘Laws of nature, natural phenomena, and abstract ideas are not
patentable.’” Alice Corp. v. CLS Bank Int’l, 573 U.S. 208, 216 (2014)
(quoting Ass’n for Molecular Pathology v. Myriad Genetics, Inc., 569 U.S.
576, 589 (2013)). To determine whether a claim falls within one of these
excluded categories, the Court has set out a two-part framework. The
framework requires us first to consider whether the claim is “directed to one
of those patent-ineligible concepts.” Alice, 573 U.S. at 217. If so, we then
examine “the elements of each claim both individually and ‘as an ordered
combination’ to determine whether the additional elements ‘transform the
nature of the claim’ into a patent-eligible application.” Alice, 573 U.S.
at 217 (quoting Mayo Collaborative Servs. v. Prometheus Labs., Inc., 566
U.S. 66, 78, 79 (2012)). That is, we examine the claims for an “inventive
concept,” “an element or combination of elements that is ‘sufficient to
ensure that the patent in practice amounts to significantly more than a patent
Appeal 2018-008697
Application 15/136,314

5
upon the [ineligible concept] itself.’” Alice, 573 U.S. at 217–18 (alteration
in original) (quoting Mayo, 566 U.S. at 72–73).
The Patent Office recently issued guidance regarding this framework.
See USPTO, 2019 Revised Patent Subject Matter Eligibility Guidance, 84
Fed. Reg. 50 (Jan. 7, 2019) (“Revised Guidance”). Under the Revised
Guidance, to decide whether a claim is “directed to” an abstract idea, we
evaluate whether the claim (1) recites subject matter falling within an
abstract idea grouping listed in the Revised Guidance and (2) fails to
integrate the recited abstract idea into a practical application. See Revised
Guidance, 84 Fed. Reg. at 51. If the claim is “directed to” an abstract idea,
as noted above, we then determine whether the claim recites an inventive
concept. The Revised Guidance explains that when making this
determination, we should consider whether the additional claim elements
add “a specific limitation or combination of limitations that are not well-
understood, routine, conventional activity in the field” or “simply append[]
well-understood, routine, conventional activities previously known to the
industry.” Revised Guidance, 84 Fed. Reg. at 56.
With these principles in mind, we turn to the Examiner’s § 101
rejection.
The Judicial Exception—Abstract Idea
The Examiner determined the claims are directed to the abstract idea
of “a mathematical process to quickly converge a derivative function f(x).”
Non-Final Act. 2; see also Ans. 2 (“Claim 1 is directed to improvements in
Appeal 2018-008697
Application 15/136,314

6
the convergence speed of solving a nonlinear optimization problem.”). For
the reasons explained below, we agree that the claims recite an abstract idea.
Limitations (1)–(7) of claim 1, quoted above, under their broadest
reasonable interpretation, recite a mathematical algorithm for solving a
nonlinear optimization problem.4 See, e.g., Spec. 1–2, 4, 8, 12. Specifically,
limitations (2), (4), (5), and (7) involve setting, determining, or updating
variables—i.e., vector g, a search direction vector, a step size, vector x—
using other variables and/or one or more mathematical algorithms—e.g., a
gradient of f(x), a parabolic approximation, a first derivative of the function
f(x), the second derivative of the function f(x), and an approximation of
another function. Limitation (6) determines if a criteria has been satisfied,
and limitations (1) and (3) merely recite executing the search specified by
the remaining limitations.
When claimed in a manner similar to the claims here, a claim reciting
a mathematical algorithm or calculation has been determined to be an
abstract idea. See Parker v. Flook, 437 U.S. 584, 594 (1978) (“[T]he
discovery of [a mathematical formula] cannot support a patent unless there is
some other inventive concept in its application.”); SAP Am., Inc. v.
InvestPic, LLC, 898 F.3d 1161, 1167 (Fed. Cir. 2018) (concluding claims for
providing statistical analysis of investment data were directed to the abstract

4 The Specification explains that solving the nonlinear optimization problem
seeks to find the “optimal values” for a machine learning system. Spec. 1–2.
“To optimize the values[,] the systems typically evaluate a function f(x) that
represents a degree of optimality or an error from the optimality of x,” where
x is a vector (i.e., a set of values to be used in the system) that is iteratively
improved “to maximize or minimize f(x) depending on what it represents.”
Id. at 2.
Appeal 2018-008697
Application 15/136,314

7
idea of “selecting certain information, analyzing it using mathematical
techniques, and reporting or displaying the results of the analysis”); Digitech
Image Techs., LLC v. Elec. for Imaging, Inc., 758 F.3d 1344, 1350 (Fed. Cir.
2014) (concluding that “a process of organizing information through
mathematical correlations and . . . not tied to a specific structure or machine”
to be abstract).
Accordingly, we conclude the claims recite a mathematical concept,
and thus, an abstract idea. Revised Guidance, 84 Fed. Reg. at 52, 53 (listing
“[m]athematical concepts—mathematical relationships, mathematical
formulas or equations, mathematical calculations” as one of the “enumerated
groupings of abstract ideas”).
Integration of the Judicial Exception into a Practical Application
If a claim recites a judicial exception, we determine whether the
recited judicial exception is integrated into a practical application of that
exception by: (a) identifying whether there are any additional elements
recited in the claim beyond the judicial exception(s); and (b) evaluating
those additional elements individually and in combination to determine
whether they integrate the exception into a practical application. If the
recited judicial exception is integrated into a practical application, the claim
is not directed to the judicial exception.
Here, claim 1 recites the additional element of “using the vector x as a
value of a parameter of the artificial neural network” (limitation (8)).
Br. 23–24 (Claims App’x). Considering claim 1 as a whole, this additional
element does not apply or use the abstract idea in a meaningful way such
that the claim as a whole is more than a drafting effort designed to
monopolize the exception. The Supreme Court guides that the “prohibition
Appeal 2018-008697
Application 15/136,314

8
against patenting abstract ideas ‘cannot be circumvented by attempting to
limit the use of the formula to a particular technological environment’ or
[by] adding ‘insignificant postsolution activity.’” Bilski, 561 U.S. at 610–11
(quoting Diamond v. Diehr, 450 U.S. 175, 191–92 (1981)).
The Specification describes the artificial neural network as a generic
component. See Spec. 44 (neural network is “control system 10,” which
includes programs and data), 46. The Specification does not address how a
vector is used as a parameter of an artificial neural network, indicating that
this is a generic function. Moreover, the Specification describes the
nonlinear optimization method as being performed by a generic computer.
See, e.g., Spec. 37 (“[T]he processing system 1 that performs nonlinear
optimization includes a computer 11 having a processing unit (CPU) 12, an
input device 13, an output device 14, and a processing program 16a and
processing data 16b that are stored in a memory 15.”), Figs. 5–7 (identifying
generic computer components); see also Non-Final Act. 3 (“[N]either the
claim nor the specification require any specific architecture not already
known and used in the art.”).
Moreover, the claim’s requirement of “using the vector x as a value of
a parameter of the artificial neural network” (limitation (8)) is the type of
extra-solution activity (i.e., in addition to the judicial exception) the courts
have determined insufficient to transform judicially excepted subject matter
into a patent-eligible application. See MPEP § 2106.05(g); see also Alice,
573 U.S. at 223 (“Stating an abstract idea while adding the words ‘apply it
with a computer’” is not sufficient to confer patent eligibility.). Indeed, in
Parker v. Flook, the claimed algorithm was an abstract idea, even though the
last step of the algorithm adjusted an alarm limit on a variable involved in
Appeal 2018-008697
Application 15/136,314

9
the catalytic chemical conversion of hydrocarbons. 437 U.S. at 594 (The
“process is unpatentable under § 101, not because it contains a mathematical
algorithm as one component, but because once that algorithm is assumed to
be within the prior art, the application, considered as a whole, contains no
patentable invention.”); see Diamond v. Diehr, 450 U.S. 175, 191 (1981)
(“[T]his principle cannot be circumvented by attempting to limit the use of
the formula to a particular technological environment.”).
Appellant argues that claim 1 is not abstract because it “improve[s]
the function of the computer.” Br. 11–12. In support, Appellant points to
Enfish and submits that claim 1 “solves a problem specific to computer-
related technology” by training a neural network “with an increased
processing speed.” Id. at 12 (citing Enfish, LLC v. Microsoft Corp., 822
F.3d 1327 (Fed. Cir. 2016)). We are not persuaded by Appellant’s
argument. The claimed self-referential table in Enfish was a specific type of
data structure designed to improve the way a computer stores and retrieves
data in memory. Enfish, 822 F.3d at 1339. Unlike the claims in Enfish,
claim 1 does not recite an improved computer (or an improved data structure
or neural network); rather, claim 1 recites an allegedly improved
mathematical algorithm for selecting a parameter of a neural network. See
id. at 1338 (“[U]nlike the claims [in Enfish] that are directed to a specific
improvement to computer functionality, the patent-ineligible claims at issue
in other cases recited use of an abstract mathematical formula on any general
purpose computer.”).
Finally, Appellant contends that “claim 1 is directed to a specific
implementation” and “is not attempting to preempt all uses of the recited
algorithm across all technologies.” Br. 14. This argument does not alter our
Appeal 2018-008697
Application 15/136,314

10
§ 101 analysis. Rather than preemption, the Federal Circuit instructs that the
Alice two-step analysis is the test for whether claims are statutory. “While
preemption may signal patent ineligible subject matter, the absence of
complete preemption does not demonstrate patent eligibility.” Ariosa
Diagnostics, Inc. v. Sequenom, Inc., 788 F.3d 1371, 1379 (Fed. Cir. 2015).
“Where a patent’s claims are deemed only to disclose patent ineligible
subject matter under the Mayo framework, as they are in this case,
preemption concerns are fully addressed and made moot.” Id.
Accordingly, even in combination with all the other recited elements,
the addition of “a neural network” does not integrate the abstract idea into a
practical application because it does not impose any meaningful limits on
practicing the abstract idea. For these reasons, we determine that claim 1
does not integrate the recited abstract idea into a practical application.
Inventive Concept
Because we agree with the Examiner that claim 1 is “directed to” an
abstract idea, we consider whether an additional element (or combination of
elements) adds a limitation that is not well-understood, routine, conventional
(“WURC”) activity in the field or whether the additional elements simply
append WURC activities previously known to the industry, specified at a
high level of generality, to the judicial exception. Revised Guidance, 84
Fed. Reg. at 23. The Examiner’s finding that an additional element (or
combination of elements) is WURC activity must be supported with a
factual determination. Id. (citing MPEP § 2106.05(d), as modified by the
Appeal 2018-008697
Application 15/136,314

11
Berkheimer Memorandum5). Whether additional elements (i.e., “a neural
network”) are WURC activity is a question of fact. See Berkheimer v. HP
Inc., 881 F.3d 1360, 1369 (Fed. Cir. 2018) (“Whether something is well-
understood, routine, and conventional to a skilled artisan . . . is a factual
determination.”).
On the record before us, Appellant has not shown that the claims on
appeal add a specific limitation beyond the judicial exception that is not
“well-understood, routine, and conventional” in the field (see MPEP
§ 2106.05(d)). The Examiner found that “neither the claim nor the
specification require any specific architecture not already known and used in
the art.” Non-Final Act. 3 (citing Spec. 38). Although Appellant argues that
the “invention is not merely the routine or conventional use of a computer”
(Br. 13), Appellant neither responds to the Examiner’s finding nor points to
a particular claimed element that does not qualify as WURC. See id.
at 13–14. In fact, Appellant’s Specification demonstrates the WURC nature
of the claimed elements because it indicates they may be implemented with
generic devices. Spec. 36–39, 41–42, 45–46, Figs. 5–7.
For these reasons, we conclude that claim 1, considered as a whole,
does not include an inventive concept.
Therefore, we sustain the Examiner’s § 101 rejection of independent
claim 1. We also sustain the Examiner’s § 101 rejection of claims 6 and 11,
for which Appellant relies on the same arguments made with respect to

5 Robert W. Bahr, Changes in Examination Procedure Pertaining to Subject
Matter Eligibility, Recent Subject Matter Eligibility Decision (Berkheimer v.
HP, Inc.) (2018) (hereinafter “Berkheimer Memorandum”).
Appeal 2018-008697
Application 15/136,314

12
claim 1. Br. 14–21. The dependent claims (i.e., claims 4, 5, 9, 10, 14, 15,
18, and 19) fall with their respective independent claim. See Br. 10.
CONCLUSION
We conclude Appellant have not demonstrated the Examiner erred in
rejecting claims 1, 4–6, 9–11, 14, 15, 18, and 19 under 35 U.S.C. § 101.
DECISION
We affirm the Examiner’s decision rejecting claims 1, 4–6, 9–11, 14,
15, 18, and 19.
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a)(1)(iv). See
37 C.F.R. § 41.50(f).
AFFIRMED