Ex Parte Eshghi

Patent Trials and Appeals Board

May 30, 2019

13873587 - (D) (P.T.A.B. May. 30, 2019)

UNITED STA TES p A TENT AND TRADEMARK OFFICE
APPLICATION NO. FILING DATE
13/873,587 04/30/2013
146568 7590
ENTIT SOFTWARE LLC
500 Westover Drive
#12603
Sanford, NC 27330
06/03/2019
FIRST NAMED INVENTOR
KaveEshghi
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
ATTORNEY DOCKET NO. CONFIRMATION NO.
83209749 9205
EXAMINER
TRAN,MAIT
ART UNIT PAPER NUMBER
2124
NOTIFICATION DATE DELIVERY MODE
06/03/2019 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):
software.ip.mail@microfocus.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte KA VE ESHGHI
Appeal2018-006632
Application 13/873,587
Technology Center 2100
Before JOSEPH L. DIXON, HUNG H. BUI, and JON M. JURGOV AN,
Administrative Patent Judges.
DIXON, Administrative Patent Judge.
DECISION ON APPEAL
Appeal2018-006632
Application 13/873,587
STATEMENT OF THE CASE
Appellant1 appeals under 35 U.S.C. § 134(a) from a rejection of
claims 1-20. We have jurisdiction under 35 U.S.C. § 6(b).
We affirm.
The claims are directed to a method, system, and computer readable
storage medium for "building a classification model using a SVM [ ( support
vector machine)] training module" by computing a mean value of a set of
training vectors, subtracting the mean value of the set of training vectors
from each training vector to obtain a set of difference vectors, applying a
hash function to each of the difference vectors to obtain a number of hashed
vectors, and applying a linear training formula to the hashed vectors to
obtain a classifier model. (Abstract.)
Claim 1, reproduced below, is illustrative of the claimed subject
matter:
1. A method comprising:
performing, by a computer system comprising a
processor:
computing a mean of a plurality of training vectors;
subtracting the mean of the plurality of training vectors
from each training vector of the plurality of training vectors to
obtain a plurality of difference vectors;
applying a hash function to each of the plurality of
difference vectors to obtain a plurality of hashed vectors;
applying linear training to the plurality of hashed vectors
to obtain a classifier model; and
1 Appellant indicates that EntIT Software LLC is the real party in interest.
(App. Br. 1.)
2
Appeal2018-006632
Application 13/873,587
classifying a sample vector of input data using the
classifier model.
(App. Br. i (Claims Appendix).)
REFERENCES
The prior art relied upon by the Examiner in rejecting the claims on
appeal is:
Matsui et al.
("Matsui")
Perronnin et al.
("Perronnin")
us 5,845,016
US 2011/0040711 Al
Dec. 1, 1998
Feb. 17, 2011
Kave Eshghi & Shyamsundar Rajaram, Locality Sensitive Hash Functions
Based on Concomitant Rank Order Statistics, Proceedings of the 14th ACM
SIGKDD Int'l Conference on Knowledge Discovery & Data Mining (KDD
'08) 221-29 (2008) ("Eshghi").
Rong-En Fan et al., LIBLINEAR: A Library for Large Linear Classification,
9 J. OF MACHINE LEARNING RESEARCH 1871-74 (2008) ("Fan")
REJECTIONS
The Examiner made the following rejections:
Claims 1-20 stand rejected under 35 U.S.C. § 101 because the
claimed invention is directed to a judicial exception (i.e., a law of nature, a
natural phenomenon, or an abstract idea) without significantly more.
Claims 1-3, 5-8, and 10-20 stand rejected under 35 U.S.C. § 103 as
being unpatentable over Perronnin in view of Matsui further in view of
Eshghi.
Claims 4 and 9 stand rejected under 35 U.S.C. § 103 as being
unpatentable over Perronnin in view of Matsui and Eshghi, and further in
view of Fan.
3
Appeal2018-006632
Application 13/873,587
ANALYSIS
35 U.S.C. § 101
An invention is patent-eligible if it claims a "new and useful process,
machine, manufacture, or composition of matter." 35 U.S.C. § 101.
However, the Supreme Court has long interpreted 35 U.S.C. § 101 to include
implicit exceptions: "'[l]aws of nature, natural phenomena, and abstract
ideas'" are not patentable. Mayo Collaborative Servs. v. Prometheus Labs.,
Inc., 566 U.S. 66, 70 (2012) (brackets in original) (citing Diamond v. Diehr,
450 U.S. 175, 185 (1981)).
In determining whether a claim falls within an excluded category, we
are guided by the Supreme Court's two-step framework, described in Mayo
and Alice. Alice Corp. v. CLS Banklnt'l, 573 U.S. 208, 217-18 (2014)
(citing Mayo, 566 U.S. at 75-77). In accordance with that framework, we
first determine what concept the claim is "directed to." See Alice, 573 U.S.
at 218-19 ("On their face, the claims before us are drawn to the concept of
intermediated settlement, i.e., the use of a third party to mitigate settlement
risk."); see also Bilski v. Kappas, 561 U.S. 593, 611 (2010) ("Claims 1 and 4
in petitioners' application explain the basic concept of hedging, or protecting
against risk.").
Concepts determined to be abstract ideas, and thus patent ineligible,
include certain methods of organizing human activity, such as fundamental
economic practices (Alice, 573 U.S. at 219-20; Bilski, 561 U.S. at 611);
mathematical formulas (Parker v. Flook, 437 U.S. 584, 594--95 (1978)); and
mental processes (Gottschalkv. Benson, 409 U.S. 63, 69 (1972)). Concepts
determined to be patent eligible include physical and chemical processes,
4
Appeal2018-006632
Application 13/873,587
such as "molding rubber products" (Diehr, 450 U.S. at 191 ); "tanning,
dyeing, making water-proof cloth, vulcanizing India rubber, smelting ores"
(id. at 183 n.7 (quoting Corning v. Burden, 56 U.S. 252, 267---68 (1853)));
and manufacturing flour (Benson, 409 U.S. at 69 ( citing Cochrane v.
Deener, 94 U.S. 780, 785 (1876))).
In Diehr, the claim at issue recited a mathematical formula, but the
Supreme Court held that "[a] claim drawn to subject matter otherwise
statutory does not become nonstatutory simply because it uses a
mathematical formula." Diehr, 450 U.S. at 187; see also id. at 191 ("We
view respondents' claims as nothing more than a process for molding rubber
products and not as an attempt to patent a mathematical formula."). Having
said that, the Supreme Court also indicated that a claim "seeking patent
protection for that formula in the abstract ... is not accorded the protection
of our patent laws, ... and this principle cannot be circumvented by
attempting to limit the use of the formula to a particular technological
environment." Id. (citing Benson and Flook); see, e.g., id. at 187 ("It is now
commonplace that an application of a law of nature or mathematical formula
to a known structure or process may well be deserving of patent
protection.").
If the claim is "directed to" an abstract idea, we tum to the second
step of the Alice and Mayo framework, where "we must examine the
elements of the claim to determine whether it contains an 'inventive
concept' sufficient to 'transform' the claimed abstract idea into a patent-
eligible application." Alice, 573 U.S. at 221 (internal quotation marks
omitted). "A claim that recites an abstract idea must include 'additional
features' to ensure 'that the [claim] is more than a drafting effort designed to
5
Appeal2018-006632
Application 13/873,587
monopolize the [abstract idea]."' Id. ( quoting Mayo, 566 U.S. at 77).
"[M]erely requir[ing] generic computer implementation[] fail[ s] to transform
that abstract idea into a patent-eligible invention." Id.
The PTO recently published revised guidance on the application of
§ 101. USPTO, 2019 Revised Patent Subject Matter Eligibility Guidance,
84 Fed. Reg. 50 (Jan. 7, 2019) ("Revised Guidance"). Under that guidance,
we first look to whether the claim recites:
(1) any judicial exceptions, including certain groupings of abstract
ideas (i.e., mathematical concepts, certain methods of organizing
human activity such as a fundamental economic practice, or mental
processes); and
(2) additional elements that integrate the judicial exception into a
practical application ("Prong Two") (see MPEP § 2106.05(a}-(c), (e}-
(h)) (9th ed. 2018).
Only if a claim (1) recites a judicial exception and (2) does not integrate that
exception into a practical application, do we then look to whether the claim:
(3) adds a specific limitation beyond the judicial exception that is not
"well-understood, routine, conventional" in the field (see MPEP
§ 2106.05(d)); or
( 4) simply appends well-understood, routine, conventional activities
previously known to the industry, specified at a high level of
generality, to the judicial exception.
See Revised Guidance.
With respect to independent claims 1, 6, and 11, Appellant presents
arguments directed to independent claim 1, and Appellant does not set forth
separate arguments for patentability of independent claims 6 and 11. (App.
6
Appeal2018-006632
Application 13/873,587
Br. 6, 13; Reply Br. 3, 6.) As a result, we select independent claim 1 as the
representative claim for the group and address Appellant's arguments
thereto. See 37 C.F.R. § 4I.37(c)(l)(iv) (2017).
Revised Guidance Step 1
Step 1 of the Revised Guidance asks whether the claimed subject
matter falls within the four statutory categories of patentable subject matter
identified by 35 U.S.C. § 101: process, machine, manufacture, or
composition of matter. See Revised Guidance. Claim 1 recites "[a]
method." Appellant does not argue the Examiner erred in concluding claim
1 falls within the four statutory categories of patentable subject matter. We
agree with the Examiner's conclusion because claim 1 falls within the
process category. (Final Act. 4.)
Revised Guidance Step 2A~Prong One
(Does the Claim Recite a Judicial Exception?)
Under Step 2A, Prong One of the Revised Guidance, we determine
whether the claims recite any judicial exceptions, including certain
groupings of abstract ideas (i.e., mathematical concepts, certain methods of
organizing human activity such as a fundamental economic practice, or
mental processes). See Revised Guidance.
The Examiner determines that independent claim 1 is directed to "a
mathematical algorithm comprising computing a mean, subtracting the
mean, applying a hash function, and applying a linear training formula to
obtain a classifier model, and classifying a sample vector," which is an
abstract idea "analogous to the other mathematical algorithms that have been
7
Appeal2018-006632
Application 13/873,587
identified by courts as being abstract ideas" in Diehr, Benson, Flook, and
Bilski. (Final Act. 5 (citing Diehr, 450 U.S. at 175; Benson, 409 U.S. at 63;
Flook, 437 U.S. at 584; Bilski, 561 U.S. at 593); Ans. 5---6.)
Appellant argues "[t]he Examiner has erred in characterizing claim 1
as directed to improving math rather than a relevant technology" and "failed
to recognize that claim 1 is directed to improving the relevant technology
(machine classification technology) or computer functionality (functionality
of a computer in perform[ing] machine classification)." (Reply Br. 4, 6.)
Appellant also argues claim 1 is not directed to an abstract idea because the
claim is similar to the claims in Enfzsh and McRO. (App. Br. 7-8, 10-11
(citing Enfish, LLC v. Microsoft Corp., 822 F.3d 1327, 1335-36 (Fed. Cir.
2016); McRO, Inc. v. Bandai Namco Games Am. Inc., 837 F.3d 1299 (Fed.
Cir. 2016)); Reply Br. 2.) Particularly, Appellant argues that similar to
McRO, "claim 1 of the present application also involves the use of 'rules,
rather than [human] artists,' to perform the classification of input data using
the trained classifier model," the claimed "rules" including "computing of
the mean, the substracting [sic] of the mean from the training vectors to
obtain difference vectors, the applying of a hash function to each of the
difference vectors to obtain hashed vectors, and the applying of linear
training to the hashed vectors to obtain a classifier model." (App. Br. 10.)
Appellant argues the claimed "rules" "improve the functioning of a
computer system (by speeding up the training and classification processes
while being accurate) when classifying data." (App. Br. 7-8, 10 (citing
Spec. ,r 10).)
Appellant's arguments are not persuasive. Although the Specification
mentions speeding up training and classification by a computer, claim 1 is
8
Appeal2018-006632
Application 13/873,587
not directed to such a computer or to improved computer functionality for
training and classification. (See Spec. ,r 10.) Claim 1 only recites a
mathematical algorithm for building a classifier model using mathematical
quantities of "training vectors," "difference vectors," and "a hash function."
Claim 1 does not recite "using concomitant rank order hash functions" to
perform "a relatively faster training and classification of input data, with all
the same accuracy as a non-linear SVM" as described in the Specification.
(See id.) Claim 1 also does not recite or require "improv[ing] computer
functionality when classifying data," "speeding up the training and
classification processes while being accurate," or "faster training and
classification of input data with high accuracy" as Appellant asserts. (See
App. Br. 7-8, 10; Reply Br. 5.) Rather, claim 1 recites mathematical
concepts and operations of an algorithm for building and using a classifier
model. (Ans. 5.) Although claims are interpreted in light of the
Specification, limitations from the Specification are not read into the claims.
In re Van Geuns, 988 F.2d 1181, 1184 (Fed. Cir. 1993).
More specifically, representative claim 1 recites the following
limitations: (1) "computing a mean of a plurality of training vectors"; (2)
"subtracting the mean of the plurality of training vectors from each training
vector of the plurality of training vectors to obtain a plurality of difference
vectors"; (3) "applying a hash function to each of the plurality of difference
vectors to obtain a plurality of hashed vectors"; (4) "applying linear training
to the plurality of hashed vectors to obtain a classifier model"; and (5)
"classifying a sample vector of input data using the classifier model." (App.
Br. i (Claims Appendix).) These limitations, under their broadest reasonable
9
Appeal2018-006632
Application 13/873,587
interpretation, recite steps that perform a mathematical algorithm for
creating a classifier model and classifying data.
Each of the recited ( 1) and (2) limitations ("computing a mean of a
plurality of training vectors" and "subtracting the mean of the plurality of
training vectors from each training vector of the plurality of training vectors
to obtain a plurality of difference vectors") carries out one or more
mathematical concepts. The Specification discloses a mean computation
involves "adding a number of training vectors together and dividing the
resulting vector by the number of training vectors used during the building
of the classification model," and determining difference vectors involves
"normalizing the feature vectors" by "[ s ]ubtracting the population mean
from all the features vectors" to obtain a number of difference vectors. (See
Spec. ,r,r 21, 30, 34.)
The recited limitation (3) ("applying a hash function to each of the
plurality of difference vectors to obtain a plurality of hashed vectors")
carries out a mathematical concept by which a "hash function may ... be
applied ... to each of the difference vectors" where "[i]n one example, the
hash function is a concomitant rank order (CRO) hash function .... [that]
maps k dimensional real vectors to sparse U dimensional vectors." (See
Spec. ,r,r 19-20, 28.)
The remaining recited limitations ( 4) and ( 5) ("applying linear
training to the plurality of hashed vectors to obtain a classifier model" and
"classifying a sample vector of input data using the classifier model")
similarly carry out mathematical concepts to (i) apply "a linear training
routine .... [that] may be a linear SVM that takes the hash vectors and
creates a classifier model," where "[ o ]ne example of a linear training routine
10
Appeal2018-006632
Application 13/873,587
is Liblinear," such that (ii) a "hashed sample vector may then be classified
... using the classifier model obtained during the building of [the]
classification model." (See Spec. ,r,r 20-22, 24, 29, 32.)
Thus, Appellant's method for building and using a classifier model recites
mathematical operations of a mathematical algorithm, similar to
mathematical concepts identified by the courts as abstract ideas. That is,
like the decimal to binary conversion in Benson, the mathematical hedging
formula in Bilski, the Arrhenius equation in Diehr, the alarm limit formula in
Flook, and the algorithm for determining abnormal conditions in Grams,
Appellant's mathematical algorithm for creating a classifier model is an
abstract mathematical concept comprising mathematical operations and
calculations (Final Act. 5). See Benson, 409 U.S. at 63---64, 67 (a "method
for converting numerical information from binary-coded decimal numbers
into pure binary numbers ... is merely a series of mathematical calculations
or mental steps, and does not constitute a patentable 'process"'); Bilski, 561
U.S. at 611-12 ("Claims 1 and 4 in petitioners' application explain the basic
concept of hedging, or protecting against risk."); Diehr, 450 U.S. at 176, 188
("Arrhenius' equation is not patentable in isolation, but when a process for
curing rubber is devised which incorporates in it a more efficient solution of
the equation, that process is at the very least not barred at the threshold by
§ 101"); Flook, 437 U.S. at 585, 594--96 (rejecting as ineligible claims
directed to the use of an algorithm to calculate an updated "alarm-limit
value" for a catalytic conversion process variable, and updating the limit
with the new value.); In re Grams, 888 F.2d 835, 837 (Fed. Cir. 1989)
("[M]athematical algorithms join the list of non-patentable subject matter
not within the scope of section 101. "). We, therefore, conclude
11
Appeal2018-006632
Application 13/873,587
representative claim 1, and grouped claims 6 and 11, recite mathematical
operations and calculations, which is one of the judicial exceptions of an
abstract idea identified in the Prong One of Step 2A of the Revised
Guidance. See Revised Guidance, 84 Fed. Reg. at 52 ( describing an abstract
idea category of "Mathematical concepts-mathematical relationships,
mathematical formulas or equations, mathematical calculations."). Thus,
representative claim 1, and grouped claims 6 and 11, not separately argued,
recites an abstract idea.
We are unpersuaded by Appellant's argument that claim 1 does not
recite an abstract idea because "none of the listed mathematical algorithms
[ of Diehr, Benson, Flook, and Bilski] are analogous to what is recited in
claim l" and "[ n Jone of the foregoing algorithms relate to classifying a
sample vector." (App. Br. 11.) Mathematical operations (such as building a
classifier model and classifying a vector as recited in the claims) involve
mathematical concepts, which are abstract ideas. As recognized by the
Examiner, the claimed limitations recite "specific mathematical
algorithms/formulas to create a classifier model which is a mathematical
model," and although "the math may be improved ... math is still an
abstract idea." (Ans. 5.)
Revised Guidance Step 2A~Prong Two
(Integration into a Practical Application)
Under Step 2A, Prong Two of the Revised Guidance, we discern no
additional element ( or combination of elements) recited in Appellant's
representative claim 1 that integrates the judicial exception into a practical
application. See Revised Guidance, 84 Fed. Reg. at 54--55 ("Prong Two").
12
Appeal2018-006632
Application 13/873,587
For example, Appellant's claimed additional elements ("a computer
system," "processor," and "storage medium") do not: (1) improve the
functioning of a computer or other technology; (2) are not applied with any
particular machine (except for a generic computer); (3) do not effect a
transformation of a particular article to a different state; and ( 4) are not
applied in any meaningful way beyond generally linking the use of the
judicial exception to a particular technological environment, such that the
claim as a whole is more than a drafting effort designed to monopolize the
exception. See MPEP §§ 2106.05(a}-(c), (e}-(h). Appellant's claim 1 does
not describe technological improvements, and is not directed to an
improvement in computer-related technology. Nor has Appellant
demonstrated that the claim recites a specific improvement to the way
computers operate or to the way computers store and retrieve data in
memory. (Ans. 5---6; see Enfish, 822 F.3d at 1336, 1339.)
The additional elements in claim 1 are broadly claimed to cover a
processor that is recited in such a general, generic, and functional manner
that the claim fails to capture how the processor's functionality is improved
when classifying data. (See Reply Br. 3.) Thus, claim 1 's hardware
elements do not evince integration into a practical application, such as a non-
abstract improvement in "the functioning of a computer system (by speeding
up the training and classification processes while being accurate) when
classifying data." (App. Br. 7-8.) Rather, Appellant's claimed invention
uses existing technology to perform a mathematical algorithm. See Bancorp
Servs., L.L.C. v. Sun Life Assurance Co. of Can. (U.S.), 687 F.3d 1266, 1278
(Fed. Cir. 2012) ("[T]he fact that the required calculations could be
performed more efficiently via a computer does not materially alter the
13
Appeal2018-006632
Application 13/873,587
patent eligibility of the claimed subject matter."); Ultramercial, Inc. v. Hulu,
LLC, 772 F.3d 709, 715-16 (Fed. Cir. 2014) ("Adding routine additional
steps such as ... use of the Internet does not transform an otherwise abstract
idea into patent-eligible subject matter.").
Further, even though claim 1 may be for a specific purpose ( e.g.,
"improving the machine classification technology" or "training and
classification processes of classifiers," (see Reply Br. 4--5)), this does not
transform the claim into statutory subject matter. As the Supreme Court has
stated, "'if a claim is directed essentially to a method of calculating, using a
mathematical formula, even if the solution is for a specific purpose, the
claimed method is nonstatutory. "' Flook, 437 U.S. at 595 ( quoting In re
Richman, 563 F.2d 1026, 1030 (CCPA 1977)).
Additionally, we note the functions performed by the processor
recited in claim 1 automate mathematical operations that can be performed
by humans using pen and paper. (Ans. 4--5.) For example, the claimed
computing a mean, subtracting the mean, applying a hash function, and
training and applying the classifier model can be performed manually,
without a computer or other machine. See CyberSource Corp. v. Retail
Decisions, Inc., 654 F.3d 1366, 1372-73 (Fed. Cir. 2011) ("[A] method that
can be performed by human thought alone is merely an abstract idea and is
not patent-eligible under§ 101."); see also In re Comiskey, 554 F.3d 967,
979 (Fed. Cir. 2009) ("[M]ental processeS--Or processes of human thinking-
standing alone are not patentable even if they have practical application.");
Benson, 409 U.S. at 67 ("Phenomena of nature ... , mental processes, and
abstract intellectual concepts are not patentable, as they are the basic tools of
scientific and technological work" (emphasis added)). Additionally, mental
14
Appeal2018-006632
Application 13/873,587
processes remain unpatentable even when automated to reduce the burden
on the user of what once could have been done with pen and paper.
CyberSource, 654 F.3d at 1375 ("That purely mental processes can be
unpatentable, even when performed by a computer, was precisely the
holding of the Supreme Court in Gottschalkv. Benson.").
We are also not persuaded that Appellant's claim 1 is similar to the
claim in McRO. (See App. Br. 7, 9-10.) Appellant's claim 1 does not
describe technological process improvements similar to McRO. In McRO,
the claim was not directed to an abstract idea because it "uses the limited
rules in a process specifically designed to achieve an improved technological
result" over "existing, manual 3-D animation techniques." See McRO, 837
F.3d at 1316. That is, McRO's claim was not directed to an abstract idea
because it "allow[ ed] computers to produce 'accurate and realistic lip
synchronization and facial expressions in animated characters' that
previously could only be produced by human animators." See id. at 1313.
In contrast, Appellant's claim 1 employs broadly claimed quantities
("training vectors," "difference vectors" and "a hash function") to obtain a
classifier model and classify a sample vector. The claim is not tied to a real-
world application or to an improved "functionality of a computer in
perform[ing] machine classification" (see Reply Br. 5); rather, the claim
merely requires a mathematical algorithm computing difference and hashed
vectors and training and using a classifier model, as discussed supra.
Additionally, the claim does not indicate how its obtained results ( e.g., the
classifier model and the classified sample vector) provide any improvement
to computer functionality or to machine classification technology. Thus,
claim 1 does not integrate the judicial exception into a practical application.
15
Appeal2018-006632
Application 13/873,587
With respect to preemption, Appellant argues the claim is not directed
to an abstract idea because "claim 1 is not attempting to cover all possible
ways of input data classification[, r ]ather, claim 1 recites a very specific set
of rules, like the rules applied in McRO." (App. Br. 11.)
As recognized by the Examiner, however, although the claimed
mathematics may be improved or specific, but "math is still an abstract
idea." (Ans. 5.) We agree with the Examiner. As the McRO court further
explicitly recognized, "the absence of complete preemption does not
demonstrate patent eligibility." See McRO, 837 F.3d at 1315 (quoting
Ariosa Diagnostics, Inc. v. Sequenom, Inc., 788 F.3d 1371, 1379 (Fed. Cir.
2015)). Furthermore, "[w]here a patent's claims are deemed only to disclose
patent ineligible subject matter" under the Alice/Mayo framework,
"preemption concerns are fully addressed and made moot." Ariosa, 788
F.3d at 1379.
For these reasons, we agree with the Examiner that claim 1, and
grouped claims 6 and 11, are directed to performing a mathematical
algorithm for creating a classifier model and classifying data, which is a
mathematical concept identified as an abstract idea in the Revised Guidance,
constituting a judicial exception to patentability and not integrating the
abstract idea into a practical application.
Revised Guidance Step 2B (Inventive Concept)
Under the Revised Guidance, only if a claim ( 1) recites a judicial
exception and (2) does not integrate that exception into a practical
application, do we then look to whether the claim:
(3) adds a specific limitation beyond the judicial exception that is not
16
Appeal2018-006632
Application 13/873,587
"well-understood, routine, conventional" in the field (see MPEP
§ 2106.05(d)); or
( 4) simply appends well-understood, routine, conventional activities
previously known to the industry, specified at a high level of
generality, to the judicial exception.
Here, the Examiner maintains "the current claims only have an
abstract idea plus a general process computer. Thus, the claims do not have
anything that have [sic] been found by courts to be significantly more" than
a judicial exception. (Ans. 7.)
Appellant contends the Examiner erred, and claim 1 recites
significantly more than a judicial exception for the reasons that: (1) claim 1
recites "subject matter that improves upon the relevant technology and
computer functionality" and "improves another technology or technical
field" particularly by "improving ... machine classification technology ...
[and] functionality of a computer in perform[ing] machine classification;"
and (2) "claim 1 is clearly inventive over the cited art." (App. Br. 13; Reply
Br. 5.)
Appellant's arguments are not persuasive. Under Federal Circuit
precedent, an "inventive concept" under Alice step 2 and the Revised
Guidance can be established by showing, for example, that the patent claims:
( 1) provide a technical solution to a technical problem
unique to the Internet, e.g., a "solution ... necessarily
rooted in computer technology in order to overcome a
problem specifically arising in the realm of computer
networks" (see DDR Holdings, LLC v. Hotels.com, L.P.,
773 F.3d 1245, 1257 (Fed. Cir. 2014));
(2) transform the abstract idea into "a particular, practical
application of that abstract idea," e.g., "installation of a
filtering tool at a specific location, remote from the end-
17
Appeal2018-006632
Application 13/873,587
users, with customizable filtering features specific to
each end user" (see Bascom Glob. Internet Servs., Inc. v.
AT&T Mobility LLC, 827 F.3d 1341, 1350, 1352 (Fed.
Cir. 2016)); or
(3) "entail[] an unconventional technological solution ([ e.g.,]
enhancing data in a distributed fashion) to a
technological problem ([e.g.,] massive record flows [that]
previously required massive databases)" and "improve
the performance of the system itself' (see Amdocs (Isr.)
Ltd. v. Openet Telecom, Inc., 841 F.3d 1288, 1300, 1302
(Fed. Cir. 2016)).
Similarly, as recognized by the Revised Guidance, an "inventive
concept" under Alice step 2 can also be evaluated based on whether an
additional element or combination of elements:
(1) "[ a ]dds a specific limitation or combination of limitations
that are not well-understood, routine, conventional
activity in the field, which is indicative that an inventive
concept may be present;" or
(2) "simply appends well-understood, routine, conventional
activities previously known to the industry, specified at a
high level of generality, to the judicial exception, which
is indicative that an inventive concept may not be
present."
See Revised Guidance, 84 Fed. Reg. at 56.
In this case, however, we find no element or combination of elements
recited in Appellant's representative claim 1 that contains any "inventive
concept" or adds anything "significantly more" to transform the abstract
concept into a patent-eligible application. Alice, 573 U.S. at 222. For
example, Appellant's claimed invention does not recite additional elements
that are rooted in computer technology, as the claims in DDR Holdings. Nor
do any of the additional elements provide any technical solution to a
technical problem as required by DDR. Appellant also has not demonstrated
18
Appeal2018-006632
Application 13/873,587
their claimed computer system and processor are able to performfunctions
that are not merely generic, as the claims in DDR. See DDR Holdings, 773
F.3d at 1258.
As discussed supra, Appellant's claim 1 merely recites mathematical
operations that can be performed manually or automated by Appellant's
generically claimed processor. (Ans. 4, 7; see also Spec. ,r 33 (providing a
"computer usable program code [that] may be provided to a processor of a
general purpose computer, special purpose computer, or other programmable
data processing apparatus" to implement the invention.).) Appellant's
mathematical algorithm does not effect a technological improvement and
does not provide any particular practical application as required by
BASCOM, or entail an unconventional technological solution to a
technological problem as required by Amdocs. Particularly, Appellant's
claim is not tied to a real-world application of classified results, or to an
actual improvement in machine classification technology or an improvement
to the "functionality of a computer in perform[ing] machine classification."
(See Reply Br. 6.) Rather, Appellant's claim uses generic technology to
perform the mathematical algorithm. As our reviewing court has observed,
"after Alice, there can remain no doubt: recitation of generic computer
limitations does not make an otherwise ineligible claim patent-eligible."
DDR Holdings, 773 F.3d at 1256 (citing Alice, 134 S. Ct. 2347, 2358
(2014)).
Appellant's additional argument that "claim 1 is clearly inventive over
the cited art" (see App. Br. 13) is not persuasive because it improperly
conflates the test for § 101 with the separate tests for § § 102 and 103. See,
e.g., Genetic Techs. Ltd. v. Merial L.L.C., 818 F.3d 1369, 1376 (Fed. Cir.
19
Appeal2018-006632
Application 13/873,587
2016) ("under the Mayo/Alice framework, a claim directed to a newly
discovered law of nature (or natural phenomenon or abstract idea) cannot
rely on the novelty of that discovery for the inventive concept necessary for
patent eligibility."). As the Supreme Court emphasizes, "[t]he 'novelty' of
any element or steps in a process, or even of the process itself, is of no
relevance in determining whether the subject matter of a claim falls within
the§ 101 categories of possibly patentable subject matter." Diehr, 450 U.S.
at 188-89 ( emphasis added). Our reviewing court further guides that
"[ e ]ligibility and novelty are separate inquiries." Two-Way Media Ltd. v.
Comcast Cable Commc 'ns, LLC, 874 F.3d 1329, 1340 (Fed. Cir. 2017); see
also Affinity Labs of Tex., LLC v. DIRECTV, LLC, 838 F.3d 1253, 1263
(Fed. Cir. 2016) (holding that "[e]ven assuming" that a particular claimed
feature was novel does not "avoid the problem of abstractness."). Although
the second step in the Alice/Mayo framework is termed a search for an
"inventive concept," the analysis is not an evaluation of novelty or non-
obviousness, but rather, a search for "an element or combination of elements
that is 'sufficient to ensure that the patent in practice amounts to
significantly more than a patent upon the [ineligible concept] itself."' Alice,
573 U.S. at 217-18 (quoting Mayo 566 U.S. 72-73). A novel and
nonobvious claim directed to a purely-abstract idea is, nonetheless, patent-
ineligible. See Mayo, 566 U.S. at 89-91.
Because Appellant's representative claim 1 is directed to a patent-
ineligible abstract concept and does not recite something "significantly
more" under the second step of the Alice analysis, we sustain the Examiner's
rejection of representative claim 1, and grouped claims 6 and 11 under
35 U.S.C. § 101.
20
Appeal2018-006632
Application 13/873,587
No separate arguments are presented for dependent claims 2-5, 7-10,
and 12-20, which fall with independent claims 1, 6, and 11.
37 C.F.R. § 4I.37(c)(l)(iv). We, therefore, sustain the rejection under
35 U.S.C. § 101 of claims 2-5, 7-10, and 12-20.
35 U.S.C. § 103
With respect to claim 1, the Examiner finds the combination of
Perronnin, Matsui, and Eshghi teaches "applying a hash function to each of
the plurality of difference vectors to obtain a plurality of hashed vectors"
and "classifying a sample vector of input data using ... [a] classifier model"
obtained by "applying linear training to the plurality of hashed vectors," as
claimed. (Final Act. 8-10; Ans. 10-11.) Particularly, the Examiner finds
Perronnin obtains a classifier model by applying linear training to a plurality
of training vectors, and Matsui obtains difference vectors by subtracting a
vectors' mean from each vector, as claimed. (Final Act. 8-9 ( citing
Perronnin ,r,r 11, 49, 64, 70, Abstract; Matsui, col. 2, 11. 38--40, col. 3, 11. 30-
32, 60-61, col. 4, 11. 43--44); Ans. 8-10.) The Examiner recognizes
Perronnin and Matsui do not teach "applying a hash function to each of the
plurality of difference vectors to obtain a plurality of hashed vectors" as
claimed, but finds Eshghi teaches "applying a hash function to vectors."
(Final Act. 8, 10 (citing Eshghi 222 (§ 2), 227 (11. 15-30), Title); Ans. 11
( citing Eshghi 224 (§ 4.1 )).) Based on these factual findings, the Examiner
concludes "[i]t would have been obvious ... to a person having ordinary
skill in the art to modify Perronnin-Matsui's teaching to include Eshghi's
teaching of 'applying a hash function to vectors' to approximate near
21
Appeal2018-006632
Application 13/873,587
neighbor problems in high dimensional spaces." (Final Act. 10 ( citing
Eshghi, Abstract).) We do not agree.
We agree with Appellant that the Examiner has shown portions of the
claimed invention, but none of Perronnin, Matsui, and Eshghi nor the
proffered combination of Perronnin, Matsui, and Eshghi, teach or suggest
applying a hash function to difference vectors to obtain hashed vectors to
which linear training is applied for obtaining a classifier model, as required
by claim 1. (App. Br. 17-19; Reply Br. 12-13.) As recognized by
Appellant, "there is absolutely no mention or hint in Eshghi of applying a
hash function to each of a plurality of difference vectors," and "[t]here is no
mention of difference vectors anywhere in Eshghi, such that a person of
ordinary skill in the art would not have been led by Eshghi to applying a
hash function to difference vectors" as recited in claim 1. (App. Br. 17-18.)
For example, portions of Eshghi cited by the Examiner (i.e., Eshghi 222
( § 2), 227 (11. 15-30), Title) do not teach or suggest "applying a hash
function to each of the plurality of difference vectors to obtain a plurality of
hashed vectors" as the Examiner asserts. (See Final Act. 8.) Rather,
Eshghi's pages 222 and 227 disclose hash functions are applied to original
data ( e.g., images in an image repository) to match similar images in the
repository. (See Eshghi 222, 227.) Particularly, Eshghi applies a hash
function to image features in multiple images and then "identif[ies]
matching hashes from different images," thereby "identifying matching
image pairs with their corresponding matching hash count" and labeling
"[i]mage pairs with matching count> a threshold T . .. as similar image
pairs." (See Eshghi 227-228.) Thus, Eshghi does not apply a hash function
to difference vectors ( obtained by subtracting a mean of a plurality of
22
Appeal2018-006632
Application 13/873,587
vectors from each vector) to obtain hashed vectors for training a classifier
model, as required in claim 1. (App. Br. 17-19.) Eshghi also does not apply
its hash function to training difference vectors as claimed, as Eshghi does
not use hash functions for training. (Id. at 18-19; Reply Br. 12.) As
explained supra, Eshghi uses a hash function to determine data similarity
between items/images, and match images.
We further find the Examiner's proffered reason to combine Eshghi
with Perronnin and Matsui to be insufficient to establish the obviousness of
claim 1 over Eshghi, Perronnin, and Matsui. Particularly, the Examiner
asserts "Eshghi is added to teach another well-known way to process
vectors, using hash function" and "[i]t would have been obvious ... to a
person having ordinary skill in the art to modify Perronnin-Matsui's
teaching to include Eshghi's teaching of 'applying a hash function to
vectors' to approximate near neighbor problems in high dimensional
spaces." (Ans. 11 (citing Eshghi 224, § 4.1); Final Act. 10 (citing Eshghi
Abstract) (emphasis added).) However, the Examiner has not explained how
or why Eshghi' s approximating near neighbor problems in high dimensional
spaces would be useful for vector training or building of a classifier model
as in Perronnin. Rather, Eshghi's hash functions for approximating near
neighbor problems in high dimensional spaces enable faster scans of items
(e.g., images) for finding similar items in a repository. (See Eshghi 221
( § 1 ), 228, Abstract.) That is, Eshghi teaches its hash functions enable faster
algorithms for determining data similarity, but Eshghi does not disclose that
a hash function's application could be useful for linear vector training or for
building a classifier model and classifying vectors, as in claim 1. (See
Eshghi 227-228, Abstract; Reply Br. 12-13.) The Examiner also has not
23
Appeal2018-006632
Application 13/873,587
explained what perceived benefit would prompt a skilled artisan to apply a
hash function (such as that ofEshghi) to Matsui's difference component
vector that is intended to provide luminance levels ( of an object) created in
photo diodes. (See Matsui, col. 8, 11. 50-55, Figs. 1-2.) As Appellant
explains:
The issue 1s not whether Eshghi discloses applying hash
functions to vectors in general, but whether Eshghi, in
combination with Perronnin and Matsui, would have provided
any teaching or hint of applying a hash function to each of a
plurality of difference vectors, where each difference vector is
produced by substracting [sic] a mean of the plurality of training
vectors from each training vector.
(App. Br. 19; Reply Br. 12-13.) We do not find the Examiner has provided
sufficient evidence to support the finding that the skilled artisan would apply
Eshghi' s hash function to Perronnin' s training vectors or to Matsui' s
luminance signals. In re Chaganti, 554 Fed. App'x 917, 922 (Fed. Cir.
2014) ("It is not enough to say that ... to do so would 'have been obvious to
one of ordinary skill.' . . . Such circular reasoning is not sufficient-more is
needed to sustain an obviousness rejection.")
In conclusion, we agree with Appellant that the Examiner does not set
forth a persuasive motivation to combine the disclosure of Eshghi with
Perronnin and Matsui. (Reply Br. 12-13; App. Br. 19.) Additionally, the
Examiner has not identified how the additional reference of Fan remedies
the above-noted deficiencies of Eshghi, Perronnin, and Matsui.
Thus, for the reasons set forth above, we do not sustain the
Examiner's rejection of independent claim 1, and claims 2-5 and 16-18
dependent therefrom. We also do not sustain the Examiner's rejection of
independent claim 6 (reciting the application of "a hash function to each of
24
Appeal2018-006632
Application 13/873,587
the plurality of difference vectors to obtain a plurality of hashed vectors" for
training a classifier model) and claim 11 (reciting "apply[ing] a hash
function to the sample difference vector to obtain a hashed sample vector"
and "classify[ing] the hashed sample vector using a classifier model trained
based on the plurality of SVM training vectors"), argued for substantially the
same reasons as claim 1, and dependent claims 7-10, 12-15, 19, and 20.
(App. Br. 19, 22.) Because the above-discussed issue is dispositive as to the
obviousness rejections of all claims on appeal, we do not reach additional
issues raised by Appellant's arguments as to the rejections of claims 2, 7, 14,
18, and 20.
CONCLUSIONS
The Examiner did not err in rejecting claims 1-20 based upon a lack
of patent eligible subject matter under 35 U.S.C. § 101, but the Examiner
erred in rejecting claims 1-20 based upon obviousness under 35 U.S.C.
§ 103.
DECISION
For the above reasons, we affirm the Examiner's patent eligibility
rejection of claims 1-20 under 35 U.S.C. § 101, but we reverse the
Examiner's obviousness rejection of claims 1-20 under 35 U.S.C. § 103.
Because we have affirmed at least one ground of rejection with
respect to each claim on appeal, we affirm the Examiner's decision rejecting
claims 1-20. See 37 C.F.R. § 4I.50(a)(l).
25
Appeal2018-006632
Application 13/873,587
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a)(l)(iv) (2016).
AFFIRMED
26