International Business Machines Corporation

Patent Trials and Appeals Board

Aug 5, 2020

14948679 - (D) (P.T.A.B. Aug. 5, 2020)

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.
14/948,679 11/23/2015 Gregoire Devauchelle GB920150167US1 8370
138363 7590 08/05/2020
IBM CORP. (WSM)
c/o WINSTEAD P.C.
P.O. BOX 131851
DALLAS, TX 75313
EXAMINER
FIGUEROA, KEVIN W
ART UNIT PAPER NUMBER
2124
NOTIFICATION DATE DELIVERY MODE
08/05/2020 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):
patdocket@winstead.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte GREGOIRE DEVAUCHELLE and
OLIVIER M. LHOMME
Appeal 2019-002689
Application 14/948,679
Technology Center 2100
Before JOSEPH L. DIXON, DAVID M. KOHUT, and
JON M. JURGOVAN, Administrative Patent Judges.
DIXON, Administrative Patent Judge.

DECISION ON APPEAL
Pursuant to 35 U.S.C. § 134(a), Appellant1 appeals from the
Examiner’s decision to reject claims 1–15.2 We have jurisdiction under
35 U.S.C. § 6(b).
We AFFIRM.

1 We use the word “Appellant” to refer to “applicant” as defined in
37 C.F.R. § 1.42(a). Appellant identifies the real party in interest as
International Business Machines Corporation. (Appeal Br. 1.)
2 Throughout this Decision we refer to the Final Rejection mailed March 21,
2018 (“Final Act.”), the Appeal Brief filed August 10, 2018 (“Appeal Br.”),
the Examiner’s Answer mailed December 27, 2018 (“Ans.”), and the Reply
Brief filed February 12, 2019 (“Reply Br.”).
Appeal 2019-002689
Application 14/948,679
2
INVENTION
The present invention relates to a “computer implemented method,
computer program product and system for generating a Bayesian network”
by determining parent sets of variables from which target variables are
functionally dependent, and generating a “Bayesian network . . . for the
variables such that one or more of the determined parent sets of variables for
the target variables are inserted into [a] graph and edges from the graph are
removed to ensure that the graph is acyclic.” (Abstract.)
Independent claim 1, reproduced below, is illustrative of the claimed
subject matter:
1. A computer implemented method for generating a
Bayesian network, the method comprising:
receiving a dataset comprising multiple instances of
multiple variables, wherein each instance defines a set of values
for the variables that form the dataset, wherein the variables carry
numerical values, text ranges or binary Boolean values relating
to a real world scenario;
selecting a target variable from the received dataset;
determining multiple parent sets of variables for the target
variable, such that, for each parent set of variables, the target
variable is functionally dependent on the respective parent set of
variables;
repeating for multiple variables of the received dataset, the
selecting of a new target variable from the received dataset and
determining of multiple parent sets of variables for the new target
variable, such that, for each parent set of variables, the new target
variable is functionally dependent on the respective parent set of
variables;
returning the determined multiple parent sets of variables
for the target variable and for the new target variable as a path
within a tree, wherein the path comprises a set of instructions,
wherein each of the set of instructions defines whether a variable
Appeal 2019-002689
Application 14/948,679
3
has been removed from a parent set or whether a variable is
required in the parent set;
analyzing the returned parent sets of variables for the
target variable and for the new target variable in order to
determine whether or not to keep the determined multiple parent
sets of variables for the target variable and for the new target
variable in a data store;
storing the determined multiple parent sets of variables for
the target variable and for the new target value in the data store
in response to determining to keep the determined multiple
parent sets of variables for the target variable and for the new
target value in the data store;
generating, by a processor, the Bayesian network
representing a real world system for the variables, the Bayesian
network comprising a directed acyclic graph of nodes and edges,
the generating including inserting one or more of the stored
determined parent sets of variables for the target variables into
the graph and removing edges from the graph to ensure that the
graph is acyclic; and
exploiting each functional dependency explicitly in order
to avoid representing functional dependencies as general
relations defined in one or more extensions to the Bayesian
network thereby saving time and memory consumption.

(Appeal Br. 18–24 (Claims Appendix).)
REJECTION
Claims 1–15 stand rejected under 35 U.S.C. § 101 as directed to
patent-ineligible subject matter. (Final Act. 2–3.)

ANALYSIS
Patent eligibility is a question of law that is reviewable de novo.
Dealertrack, Inc. v. Huber, 674 F.3d 1315, 1333 (Fed. Cir. 2012).
Appeal 2019-002689
Application 14/948,679
4
Accordingly, we review the Examiner’s § 101 determinations concerning
patent eligibility under this standard.
Patentable subject matter is defined by 35 U.S.C. § 101, as follows:
[w]hoever invents or discovers any new and useful process,
machine, manufacture, or composition of matter, or any new and
useful improvement thereof, may obtain a patent therefor, subject
to the conditions and requirements of this title.
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. Pty. Ltd. v. CLS Bank Int’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. Kappos, 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 (Gottschalk v. Benson, 409 U.S. 63, 67 (1972)). Concepts
Appeal 2019-002689
Application 14/948,679
5
determined to be patent eligible include physical and chemical processes,
such as “molding rubber products” (Diehr, 450 U.S. at 191); “tanning,
dyeing, making water-proof cloth, vulcanizing India rubber, smelting ores”
(id. at 182 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 turn 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 citation
omitted). “A claim that recites an abstract idea must include ‘additional
Appeal 2019-002689
Application 14/948,679
6
features’ to ensure ‘that the [claim] is more than a drafting effort designed to
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 published revised guidance on the application of § 101.
USPTO’s Memorandum, 2019 REVISED PATENT SUBJECT MATTER
ELIGIBILITY GUIDANCE, 84 Fed. Reg. 50 (January 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 (see MANUAL OF PATENT EXAMINING PROCEDURE
(MPEP) § 2106.05(a)–(c), (e)–(h) (9th ed., Rev. 08.2017, 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.
Appellant argues independent claims 1, 6, and 11 together. (See
Appeal Br. 6, 8, 11–12, 16; Reply Br. 2, 4, 6, 10–13.) 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. § 41.37(c)(1)(iv).

Appeal 2019-002689
Application 14/948,679
7
Step 1 of the Revised Guidance
Independent claim 1, as a “method” claim, recites one of the
enumerated categories of statutory subject matter in 35 U.S.C. § 101,
namely, a process. The issue before us is whether this claim is directed to a
judicial exception without significantly more.

Alice/Mayo—Step 1 (Abstract Idea)
Step 2A–Prongs 1 and 2 identified in the Revised Guidance
Step 2A, Prong 1 of the Revised Guidance
The first Prong of Step 2A under the Revised Guidance is to
determine whether the claim recites a judicial exception including
(a) mathematical concepts; (b) certain methods of organizing human
activity; and (c) mental processes. Revised Guidance, 84 Fed. Reg. at 51–
52.
The Examiner determines that claim 1 is “directed to the construction
of a Bayesian network” which is “similar to know[n] abstract ideas such as
‘Organizing information through mathematical correlations (Digitech).’”
(Final Act. 2 (citing Digitech Image Techs., LLC v. Elecs. for Imaging, Inc.,
758 F.3d 1344 (Fed. Cir. 2014)); Ans. 3–4, 8.)
Appellant argues the Examiner is generalizing the claim limitations at
a high level of abstraction, ignoring that claim 1 generates “a Bayesian
network representing a real world system” and “improves computer
capabilities by reducing memory consumption.” (Appeal Br. 8, 11.) As
such, Appellant argues claim 1 is not directed to an abstract idea but is
“directed to exploiting the functional dependencies between variables to
build more accurate Bayesian networks” and improving
Appeal 2019-002689
Application 14/948,679
8
the learning algorithm for Bayesian networks . . . in several
ways: firstly in that the method pushes further the limits of what
can be learned and secondly that the method improves the
efficiency of the learning algorithm by exploiting each functional
dependency explicitly in order to avoid representing the
functional dependencies as general relations defined in one or
more extensions to the Bayesian network, which saves time and
memory consumption.

(Appeal Br. 6–7; see also Reply Br. 7–8.) Appellant argues that Examiner’s
citation to Digitech does not support the Examiner’s assertion that claim 1 is
directed to an abstract idea, because “Appellant’s claimed invention is not
simply employing mathematical algorithms to manipulate existing
information to generate additional information” and “is not simply
combining data sets into a single data set to organize the information into a
new form.” (Appeal Br. 7–8; Reply Br. 8.)
Appellant’s arguments do not persuade us that the claims do not recite
an abstract idea, and we concur with the Examiner’s conclusion that the
claims recite an abstract idea. (Final Act. 2; Ans. 3–4, 8.)
Under its broadest reasonable interpretation, claim 1 recites
mathematical concepts and operations of an algorithm for generating a
Bayesian network, as well as an abstract mental process of generating a
Bayesian network tree. In particular, claim 1 recites an abstract mental
process of building a Bayesian network by: collecting/gathering information
(“receiving a dataset comprising multiple instances of multiple variables,
wherein each instance defines a set of values for the variables that form the
dataset, wherein the variables carry numerical values, text ranges or binary
Boolean values relating to a real world scenario” and “selecting a target
variable from the received dataset” as recited in claim 1); and analyzing the
Appeal 2019-002689
Application 14/948,679
9
information and providing results of the collection and analysis (claim 1’s
“determining multiple parent sets of variables for the target variable,”
“repeating for multiple variables of the received dataset, the selecting of a
new target variable from the received dataset and determining of multiple
parent sets of variables for the new target variable,” “returning the
determined multiple parent sets of variables for the target variable and for
the new target variable as a path within a tree” wherein “the path comprises
a set of instructions” defining “whether a variable has been removed from a
parent set or whether a variable is required in the parent set,” “analyzing the
returned parent sets of variables for the target variable and for the new target
variable in order to determine whether or not to keep the determined
multiple parent sets of variables,” “storing the determined multiple parent
sets of variables for the target variable and for the new target value,”
“generating . . . the Bayesian network representing a real world system for
the variables” by “inserting one or more of the stored determined parent sets
of variables for the target variables into the graph and removing edges from
the graph to ensure that the graph is acyclic,” and “exploiting each
functional dependency explicitly in order to avoid representing functional
dependencies as general relations defined in one or more extensions to the
Bayesian network”). (Appeal Br. 18–19 (claim 1).)
That is, although claim 1 recites a “computer implemented” method
and the use of a “processor,” the underlying steps recited in the claim are
acts that could be performed mentally and by pen and paper, without the use
of a computer or any other machine. For example, a person could manually
select (e.g., write down with pen and paper) a target variable (e.g., a “force”)
from a dataset, mentally determine multiple parent sets (e.g., a parent set
Appeal 2019-002689
Application 14/948,679
10
including “mass” and “acceleration,” and a second parent set including
“mass,” “velocity,” and “time”) such that the target variable (“force”) is
functionally dependent on each parent set, and then repeat this procedure for
other target variables in the set. Appellant’s Specification explains that
“[w]hen the value of a variable X is exactly determined by the values taken
by a set of variables S, then X is said to be functionally dependent on S.”
(See Spec. ¶ 18 (emphasis added).) A person could appreciate (mentally, or
by writing it down) that particular sets of parent variables (such as, e.g., a
set of “mass” and “acceleration,” or a set of “mass,” “velocity,” and “time”)
enable a determination of values for a target variable (such as “force”) in
accordance with a mathematical relationship or physical law (such as
Newton’s Second Law). (See Spec. ¶¶ 20 (explaining that functional
dependencies, e.g., between target variables and their parent variables “may
occur for many different reasons, but the main reason is that the data defined
by the variables respects some mathematical or physical law(s), and a formal
model is not known or not explicitly expressed within the dataset”), 32
(describing variables that “can carry numerical values (1–100 for example),
text ranges (low, medium, high) or binary Boolean values (true or false),
depending upon the nature of the thing that is being captured by the
variable”).) Claim 1’s next steps of “returning the determined multiple
parent sets of variables for the target variable and for the new target variable
as a path within a tree,” “analyzing the returned parent sets of variables,”
and “storing,” are performable in the human mind by realizing that particular
parent variables (“mass” and “acceleration,” for example) are required to
determine the value of a target variable (e.g., “force”), and, therefore, values
for “mass” and “acceleration” should be stored to enable determination of a
Appeal 2019-002689
Application 14/948,679
11
resultant “force.” Claim 1’s steps of generating a Bayesian network and
exploiting each functional dependency explicitly are performable by pen and
paper, e.g., by drawing a tree or graph illustrating dependencies between
target variables and their determining parent variables. The broad
limitations in claim 1 implement on a computer the types of mental analyses
people (e.g., students) perform when learning and analyzing mathematical
relationships, and physical laws expressing dependencies between physical
quantities.
Our reviewing court has concluded that mental processes include
similar concepts of collecting, manipulating, and providing, data. See
Intellectual Ventures I LLC v. Capital One Fin. Corp., 850 F.3d 1332, 1340
(Fed. Cir. 2017) (the Federal Circuit held “the concept of . . . collecting
data, . . . recognizing certain data within the collected data set,
and . . . storing that recognized data in a memory” ineligible); see also
Content Extraction & Transmission LLC v. Wells Fargo Bank, Nat’l Ass’n,
776 F.3d 1343, 1347 (Fed. Cir. 2014) (claims are drawn to the basic concept
of data recognition and storage); Elec. Power Grp., LLC v. Alstom S.A., 830
F.3d 1350, 1353 (Fed. Cir. 2016) (merely selecting information, by content
or source, for collection, analysis, and display does nothing significant to
differentiate a process from ordinary mental processes); CyberSource Corp.
v. Retail Decisions, Inc., 654 F.3d 1366, 1375 (Fed. Cir. 2011) (purely
mental processes can be unpatentable, even when performed by a computer).
We also agree with the Examiner that claim 1 recites operations that
manipulate data through mathematical correlations, similar to mathematical
concepts identified by the courts as abstract ideas. (Final Act. 2; Ans. 3–4,
8.) Claim 1’s steps carry out mathematical concepts of determining
Appeal 2019-002689
Application 14/948,679
12
conditional dependencies between variables, determining whether variables
are correlated with each another, and organizing variables by
interdependency and correlations in an acyclic graph of nodes and edges.
(Id.) For example, claim 1’s “determining” (multiple parent sets of
variables), “returning,” and “analyzing” steps determine parent variables that
predict a value of a target variable through a dependency expressed by, e.g.,
a physical or mathematical law. (See Spec. ¶ 20 (“Functional dependencies
between the different variables expressed in the dataset 18 may occur for
many different reasons, but the main reason is that the data defined by the
variables respects some mathematical or physical law(s)”).) Like the
decimal to binary conversion in Benson the alarm limit formula in Flook,
and the image data processing in Digitech, Appellant’s mathematical
algorithm for organizing variables in a Bayesian network is an abstract
mathematical concept comprising mathematical operations. (Final Act. 2;
Ans. 3–4, 8.) 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’”); 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); and Digitech, 758 F.3d
at 1350 (“Without additional limitations, a process that employs
mathematical algorithms to manipulate existing information to generate
additional information is not patent eligible.”).
We are unpersuaded by Appellant’s arguments that claim 1 does not
recite an abstract idea because the claim “is not simply manipulating existing
Appeal 2019-002689
Application 14/948,679
13
information into a new form.” (Reply Br. 8–9; see also Appeal Br. 7–8.)
Appellant’s arguments are not commensurate with the scope of claim 1. For
example, Appellant argues “functional dependencies produced by the parent
set generator are all minimal” and “multiple parent sets of variables . . . [are]
determined for the target variable without manipulating existing
information”—features not supported by corresponding language in claim 1.
(See Reply Br. 9.) Claim 1 does not require producing only minimal
functional dependencies, and does not exclude the use of other (unclaimed)
information or knowledge at the determining steps (of determining parent
variables).
Thus, we conclude that representative claim 1 recites abstract ideas (a
mental process of analyzing and determining interdependencies between
variables, and a mathematical concept of organizing information through
mathematical correlations) identified in the Revised Guidance. See Revised
Guidance, 84 Fed. Reg. at 51–52 (describing abstract idea categories of
“Mathematical concepts” and “Mental processes”); see also RecogniCorp,
LLC v. Nintendo Co., Ltd., 855 F.3d 1322, 1327 (Fed. Cir. 2017) (“Adding
one abstract idea . . . to another abstract idea . . . does not render the claim
non-abstract.”).
We now turn to Step 2A, Prong 2, of the Revised Guidance to
determine whether the abstract idea is integrated into a practical application.
See Revised Guidance, 84 Fed. Reg. at 54–55.
Step 2A, Prong 2 of the Revised Guidance
Under Revised Step 2A, Prong Two of the Revised Guidance, we
discern no additional element (or combination of elements) recited in
Appellant’s claim 1 that may have integrated the judicial exception into a
Appeal 2019-002689
Application 14/948,679
14
practical application. See Revised Guidance, 84 Fed. Reg. at 54–55. For
example, Appellant’s claimed additional elements (e.g., a “computer
implemented” method, “a data store,” and “a processor”) do not: (1) improve
the functioning of a computer or other technology; (2) are not applied with
any particular machine (except for generic computing elements); (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); Ans. 8.) Rather,
Appellant’s claimed computing elements are configured to perform real-
world functions and operations that automate actions and operations that can
be performed in the human mind and with pen and paper, adding nothing of
substance to the underlying abstract idea. (Ans. 6, 8.) It is clear from the
claims and the Specification (describing “a processor of a general purpose
computer, special purpose computer, or other programmable data processing
apparatus to produce a machine,” and “[a] non-exhaustive list of . . .
examples of the computer readable storage medium includ[ing] the
following: a portable computer diskette, a hard disk, a random access
memory (RAM), a read-only memory (ROM), an erasable programmable
read-only memory (EPROM or Flash memory), a static random access
memory (SRAM)”), the claimed computing elements require no
improvements to computer technology that Appellant claims to have
invented. (See Spec. ¶¶ 36, 40.) Claim 1 does not recite a specific
improvement to the way computers operate; rather, the broad claim language
recites generic “processor” and “computer implemented” limitations and
Appeal 2019-002689
Application 14/948,679
15
generic automation of operations performable in the human mind or with
pen and paper. Thus, the claim’s limitations are not indicative of
“integration into a practical application.” See Revised Guidance, 84 Fed.
Reg. at 54–55. Rather, the data store and processor are readily available
computing elements using their already available basic functions as tools in
executing the claimed method for generating a Bayesian network. See SAP
Am., Inc. v. InvestPic LLC, 898 F. 3d 1161 (Fed. Cir. 2018).
Appellant argues claim 1 is not directed to an abstract idea, as it is
directed to a practical application (as required by the Revised Guidance) for
the reasons that: (i) claim 1’s combination of steps “reflect an improvement
in the functioning of the computer or an improvement to another technology
or technical field” by “saving time and memory consumption by exploiting
each functional dependency explicitly in order to avoid representing the
functional dependencies as general relations defined in one or more
extensions to the Bayesian network” (Appeal Br. 9; see also Reply Br. 3–5
(citing Spec. ¶ 20)); (ii) similar to Enfish, “the claimed invention improves
computer capabilities by reducing memory consumption” (Reply Br. 6–7
(citing Enfish, LLC v. Microsoft Corp., 822 F.3d 1327, 1335–37 (Fed. Cir.
2016)); Appeal Br. 9, 11–12); and (iii) with claim 1, “the learning algorithm
is improved, which improves the functioning of the computer” by “saving
time and memory consumption” (Reply Br. 10–11; Appeal Br. 12).
We remain unpersuaded by Appellant’s arguments. An
“improvement in computer capabilities” as Appellant argues (see Reply Br.
4–6) is not evidenced by claim 1. Although claim 1 recites “exploiting each
functional dependency explicitly” thereby “saving time and memory
consumption,” the claim does not specify what “exploiting each functional
Appeal 2019-002689
Application 14/948,679
16
dependency explicitly” entails. (See Appeal Br. 18–19 (claim 1).) Claim 1
recites that each functional dependency is explicitly exploited “in order to
avoid representing functional dependencies as general relations defined in
one or more extensions to the Bayesian network,” but the claim does not
specify what representing functional dependencies as general relations
defined in one or more extensions to the Bayesian network would entail, or
why “representing functional dependencies as general relations” would
increase processing time or memory consumption. (See id.)
Appellant’s Specification describes variables respecting mathematical
or physical laws for which “a formal model is not known or not explicitly
expressed within the [variables’] dataset,” and provides that
[e]xplicitly handling these dependencies improves a learning
algorithm for Bayesian networks in several ways: firstly in that
the method pushes further the limits of what can be learned and
secondly that the method improves the efficiency of the learning
algorithm by exploiting each functional dependency explicitly in
order to avoid representing the functional dependencies as
general relations defined in one or more extensions to the
Bayesian network, which saves time and memory consumption.

(See Spec. ¶ 20; see also Reply Br. 6 (citing Spec. ¶ 20).) Appellant’s claim
1, however, does not require handling dependencies of variables for which
“a formal model is not known or not explicitly expressed within the dataset
[of variables].” (See id. (emphasis added).) The broad language of claim 1
includes handling functional dependencies for variables obeying known laws
(such as, e.g., Newton’s Second Law). And, for such easily modeled
variables, it is not clear how claim 1 would “push[] further the limits of what
can be learned,” or how claim 1 would “improve[] the efficiency of the
learning algorithm,” or “save[] time and memory consumption.” (See id.)
Appeal 2019-002689
Application 14/948,679
17
Appellant also discusses a problem of “a functional dependency . . . in
the data such that the size of the Cartesian product of the possible values for
a node and the node’s parents is behind the limits inherent to the learning
method” such that “this dependency cannot be learned. . . . [and] the learned
Bayesian network will not perfectly fit the distribution given by the data.”
(Reply Br. 5–6 (citing Spec. ¶¶ 4, 19–20); Appeal Br. 10.) Appellant asserts
the “claimed invention addresses such a need by exploiting functional
dependencies between variables in order to build a more accurate Bayesian
network.” (Id.) Appellant’s argument is unpersuasive because claim 1 does
not require learning a complex functional dependency, or processing
variables/nodes with functional dependencies for which “the size of the
Cartesian product of the possible values for a node and the node’s parents is
behind the limits inherent to the learning method.” (See Reply Br. 5
(emphasis added).) Claim 1 merely requires target variables that “carry
numerical values, text ranges or binary Boolean values relating to a real
world scenario,” and “multiple parent sets of variables for the target
variable, such that . . . the target variable is functionally dependent on the
respective parent set of variables.” (See Appeal Br. 18–19 (claim 1).)
We are also unpersuaded by Appellant’s arguments regarding Enfish.
(See Reply Br. 6–7.) Enfish’s data storage and retrieval method and system
recite a “self-referential table . . . [for a computer database] [which] is a
specific type of data structure designed to improve the way a computer
stores and retrieves data in memory.” Enfish, 822 F.3d at 1336, 1339. In
contrast to Enfish, Appellant’s Specification and claim 1 do not describe
technological improvements, or a specific improvement to the way
computers store and retrieve data in memory. (See Enfish, 822 F.3d at 1336,
Appeal 2019-002689
Application 14/948,679
18
1339; Ans. 5–6.) Rather, Appellant’s Specification and claim 1 describe “a
method, system and computer program product for generating a Bayesian
network,” and more particularly, a method and system for “improv[ing] a
learning algorithm for Bayesian networks” and “improved generation of a
Bayesian network.” (See Spec. ¶¶ 1, 20; Title (capitalization altered).) As
the Examiner notes, improving and optimizing the learning algorithm for a
Bayesian network is “optimizing the abstract idea of learning data” which
“is still a[n] abstract idea.” (Ans. 5–6.) The results-based-functional
language in claim 1 (“thereby saving time and memory consumption”) fails
to capture how the claimed method would improve the way computers store,
process, or retrieve data in memory. SiRF Tech., Inc. v. Int’l Trade
Comm’n, 601 F.3d 1319, 1333 (Fed. Cir. 2010) (“In order for the addition of
a machine to impose a meaningful limit on the scope of a claim, it must play
a significant part in permitting the claimed method to be performed, rather
than function solely as an obvious mechanism for permitting a solution to be
achieved more quickly, i.e., through the utilization of a computer for
performing calculations.”) As the Examiner observes, claim 1 does not
reflect an impact on memory usage or an impact on the functions of the
computer itself, and the “claim does not improve computer capabilities, [as]
it is about the generation of a Bayesian network.” (Ans. 5–6; see Intellectual
Ventures I, 850 F.3d at 1342 (explaining that “[o]ur law demands more”
than claim language that “provides only a result-oriented solution, with
insufficient detail for how a computer accomplishes it”).)
Appellant’s additional argument that “the learning algorithm is
improved, which improves the functioning of the computer” by “saving time
and memory consumption” (see Reply Br. 10–11 and Appeal Br. 12) is not
Appeal 2019-002689
Application 14/948,679
19
persuasive. The claimed “generating, by a processor, the Bayesian network”
comprising “a directed acyclic graph of nodes and edges” including a (“one
or more”) stored parent set of variables, and the claimed “exploiting each
functional dependency explicitly” fail to capture how the processor’s
functionality is improved when generating the Bayesian network.
Appellant’s claimed invention uses existing technology to perform a
mathematical algorithm (generate a Bayesian network) using steps that are
also readily performable in the human mind, and “any improvement to the
learning model [for Bayesian networks] is improving the abstract idea, not
improving computer’s capabilities.” (Ans. 5–6.) 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 patent eligibility of
the claimed subject matter.”).
We are also unpersuaded by Appellant’s arguments that claim 1 is not
abstract because the claimed “variables carry . . . values relating to a real
world scenario (e.g., healthcare scenario[)]” and “a Bayesian network
representing a real world system is generated,” such that “Appellant’s
claimed invention is directed to a real world output and it does affect the
‘tangible world.’” (Reply Br. 10; see also Appeal Br. 8.) Claim 1 merely
recites variables carrying values “relating to” a real world scenario and a
Bayesian network “representing a real world system for the variables,” but
the claim does not recite or require producing a “real world output” or
applying the Bayesian network to healthcare as Appellant argues (see id.).
In addition, the Supreme Court emphasized in Bilski that “although the
machine-or-transformation test is not the only test for patentability, this by
Appeal 2019-002689
Application 14/948,679
20
no means indicates that anything which produces a ‘useful, concrete, and
tangible result,’ State Street Bank & Trust Co. v. Signature Financial Group,
Inc., 149 F.3d 1368, 1373 (C.A. Fed. 1998), is patentable.” Bilski, 561 U.S.
at 658–660 (Breyer, J., concurring). That is, “not every claim that recites
concrete, tangible components escapes the reach of the abstract-idea
inquiry.” In re TLI Commc’ns LLC Patent Litig., 823 F.3d 607, 611
(Fed. Cir. 2016).
For these reasons, we determine claim 1, and grouped claims 6 and
11, do not integrate a judicial exception into a practical application, and
are directed to a judicial exception (namely, a mental process of analyzing
and determining interdependencies between variables, and a mathematical
concept of organizing information through mathematical correlations)
identified as an abstract idea in the Revised Guidance. Therefore, we
proceed to Step 2B, The Inventive Concept.

Alice/Mayo—Step 2 (Inventive Concept)
Step 2B identified in the Revised Guidance

As recognized by the Revised Guidance, an “inventive concept” under
Alice step 2 can 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.”
Appeal 2019-002689
Application 14/948,679
21
See Revised Guidance, 84 Fed. Reg. at 56.
We now determine whether representative independent claim 1 recites
any elements additional to the abstract idea that are not well-understood,
routine, or conventional. See MPEP § 2106.05(d). We are unable to
identify any.
The Examiner asserts,
The claim(s) does/do not include additional elements that are
sufficient to amount to significantly more than the judicial
exception because . . . the additional computer elements, which
are recited at a high level of generality, provide conventional
computer functions that do not add meaningful limits to
practicing the abstract idea. The computer elements recited
include the dataset of variables, the target variable, and the
Bayesian network which represent data in the generic computing
system/medium. There is no real world output or improvement
to a technology or field. . . . [T]he claim is simply reciting the
generation of a Bayesian network and making internal
adjustments to the data in a generic computer.

(Final Act. 2–3.)
Appellant argues claim 1 recites “significantly more” because: (i) the
Examiner “has not provided any evidence that any of the claim limitations of
Appellant’s claimed invention are routine, conventional and well-understood
that were previously engaged in by those in the field of the present
invention” and “has failed to perform [a factual determination] . . . as to
whether an element (or combination of elements) is widely prevalent or in
common use” (Appeal Br. 14; Reply Br. 11–12 (citing Berkheimer v. HP
Inc., 881 F.3d 1360 (Fed. Cir. 2018))); (ii) claim 1’s limitations “are used to
generate a Bayesian network in an unconventional way” and are not taught
by prior art, “thereby illustrating that the limitations of Appellant’s claimed
invention cannot be routine, conventional and well-understood” (Appeal Br.
Appeal 2019-002689
Application 14/948,679
22
14; Reply Br. 12–13); and (iii) the claim is patent-eligible for reasons similar
to those discussed in BASCOM (Appeal Br. 15 (citing BASCOM Glob.
Internet Servs., Inc. v. AT&T Mobility LLC, 827 F.3d 1341, 1353
(Fed. Cir. 2016))).
Appellant’s arguments are not persuasive. Particularly, we are not
persuaded that the Examiner has failed to produce factual support or
evidence that claim 1 is routine and conventional. The Examiner has
identified conventional techniques for generating Bayesian networks. (Ans.
7.)3 In addition, Appellant’s claim 1 employs generic computer elements
performing generic computer functions, as evidenced by the Specification.
(See Spec. ¶¶ 36, 38, 40; Ans. 6.) The claimed storage and data analysis
steps—that automate manually performable steps—are performed by basic
computer functions, previously known to the industry. (See Final Act. 3;
Ans. 7; see also Versata Dev. Group, Inc. v. SAP Am., Inc., 793 F.3d 1306,
1334 (Fed. Cir. 2015) (receiving, storing, retrieving, sorting, and eliminating
information is well known); In re Katz Interactive Call Processing Patent
Litig., 639 F.3d 1303, 1316 (Fed. Cir. 2011) (“Absent a possible narrower
construction of the terms ‘processing,’ ‘receiving,’ and ‘storing,’ . . . those
functions can be achieved by any general purpose computer without special
programming.”).) “[T]he use of generic computer elements like a
microprocessor or user interface” to perform conventional computer
functions “do not alone transform an otherwise abstract idea into patent-
eligible subject matter.” FairWarning IP, LLC v. Iatric Sys., Inc., 839 F.3d

3 For example, the Examiner has identified Hulten (US 2006/0112190 A1,
published May 25, 2006) as disclosing techniques for constructing Bayesian
networks. (See Hulten ¶¶ 5, 62.)
Appeal 2019-002689
Application 14/948,679
23
1089, 1096 (Fed. Cir. 2016) (citing DDR Holdings, LLC, v. Hotels.com,
L.P., 773 F.3d 1245, 1256 (Fed. Cir. 2014)).
Additionally, Appellant’s abstract idea (of a mental process of
analyzing and determining interdependencies between variables, and
mathematical concept of organizing information through mathematical
correlations) applied to generic computing infrastructure, does not provide
any particular practical application as required by BASCOM. (Ans. 7–8; see
BASCOM, 827 F.3d at 1352, 1350.) For example, BASCOM’s patent-
eligible ordered combination of claim limitations contains an “inventive
concept [that] harnesses [a] . . . technical feature of network technology in a
filtering system by associating individual accounts with their own filtering
scheme and elements while locating the filtering system on an ISP [(Internet
Service Provider)] server.” See BASCOM, 827 F.3d at 1350. BASCOM’s
claimed ordered combination “improve[s] the performance of the computer
system itself” with a “technology-based solution . . . to filter content on the
Internet that overcomes existing problems with other Internet filtering
systems.” See BASCOM, 827 F.3d at 1351–52 (internal citation omitted).
Appellant’s abstract idea using generically-claimed computing elements
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. See Amdocs (Isr.) Ltd. v.
Openet Telecom, Inc., 841 F.3d 1288, 1300, 1302 (Fed. Cir. 2016). Claim
1’s elements, considered as an ordered combination, do not improve the
functioning of a computer itself, or affect an improvement in another
technology or technical field. (Ans. 6, 8.) Instead, claim 1 amounts to
nothing significantly more than an instruction to apply the abstract idea
Appeal 2019-002689
Application 14/948,679
24
using a generic processor. That is not enough to transform an abstract idea
into a patent-eligible invention.
Appellant also argues claim 1 recites “significantly more” because the
claimed limitations are not taught by prior art, “thereby illustrating that the
limitations of Appellant’s claimed invention cannot be routine, conventional
and well-understood.” (Appeal Br. 14; Reply Br. 12–13.) This argument
improperly conflates the test for 35 U.S.C. § 101 with the separate tests for
35 U.S.C. §§ 102 and 103. (See, e.g., Genetic Techs. Ltd. v. Merial L.L.C.,
818 F.3d 1369, 1376 (Fed. Cir. 2016) (“[U]nder 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.”); see also Ans. 6–7.) 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). Thus, 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, and grouped claims 6 and
11 are directed to a patent-ineligible abstract concept and do not recite an
“inventive concept” under the second step of the Alice analysis, we sustain
the Examiner’s § 101 rejection of independent claims 1, 6, and 11, and
dependent claims 2–5, 7–10, and 12–15, not separately argued.
DECISION SUMMARY
The Examiner’s rejection of claims 1–15 under 35 U.S.C. § 101 is
AFFIRMED.
Appeal 2019-002689
Application 14/948,679
25
In summary:

Claims
Rejected
35 U.S.C.
§
Reference(s)/
Basis
Affirmed Reversed
1–15 101 Eligibility 1–15
FINALITY AND RESPONSE
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