Linkedin Corporation

Patent Trials and Appeals Board

Oct 8, 2020

2020003327 (P.T.A.B. Oct. 8, 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.
14/814,397 07/30/2015 Aman Grover 3080.E15US1 6392
45839 7590 10/08/2020
Schwegman Lundberg & Woessner / LinkedIn/Microsoft
PO BOX 2938
MINNEAPOLIS, MN 55402
EXAMINER
OFORI-AWUAH, MAAME
ART UNIT PAPER NUMBER
3629
NOTIFICATION DATE DELIVERY MODE
10/08/2020 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):
USPTO@slwip.com
slw@blackhillsip.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
____________________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
____________________

Ex parte AMAN GROVER, SIYU YOU,
KRISHNARAM KENTHAPADI, PARUL JAIN,
FEDOR VLADIMIROVICH BORISYUK,
CHRISTOPHER MATTHER DEGIERE,
and SONGTAO GUO
__________________

Appeal 2020-003327
Application 14/814,397
Technology Center 3600
____________________

Before JOSEPH A. FISCHETTI, BIBHU R. MOHANTY, and
JAMES P. CALVE, Administrative Patent Judges.

CALVE, Administrative Patent Judge.

DECISION ON APPEAL

STATEMENT OF THE CASE
Pursuant to 35 U.S.C. § 134(a), Appellant1 appeals from the decision
of the Examiner to reject claims 1–20. We have jurisdiction under 35 U.S.C.
§ 6(b).
We AFFIRM.

1 “Appellant” refers to “applicant” as defined in 37 C.F.R. § 1.42. Appellant
identifies Microsoft Technology Licensing LLC as the real party in interest.
Appeal Br. 2.
Appeal 2020-003327
Application 14/814,397

2
CLAIMED SUBJECT MATTER
Claims 1, 8, and 16 are independent. Claim 1 is reproduced below.
1. A computer-implemented method comprising:
using a machine learning algorithm to learn weights,
wherein the learning involves optimizing weights assigned to
features of member profiles, based on sample member profiles
of successful candidates, wherein a successful candidate is a
member who previously applied for a particular job and
obtained the particular job;
receiving, from a jobs database, metadata pertaining to
job listing data corresponding to a job opening, the metadata
including one or more requirements for the job opening;
obtaining, from a member database distinct from the jobs
database, a member profile for a first member of a social
networking service;
obtaining, from the member database, a plurality of
member profiles for members of the social networking service
other than the first member, the plurality of member profiles
corresponding to members who have applied for and obtained
job offers for jobs similar to the job opening;
determining, for each of the plurality of member profiles,
a date on which the corresponding member applied for the job
similar to the job opening;
obtaining, for each of the plurality of member profiles, a
version of the corresponding member profile as it was on the
date on which the corresponding member applied for the job
similar to the job opening, each of the versions of the
corresponding member profiles including a vector of features of
the corresponding member;
applying, using a computer processor, one or more of the
weights to the vectors for the plurality of member profiles to
obtain a vector of features representing a perfect candidate for
the job opening;
obtaining a vector of features of the first member from
the member profile for the first member;
calculating a distance for the first member with respect to
the perfect candidate for the job opening by comparing the
Appeal 2020-003327
Application 14/814,397

3
vector of features representing the perfect candidate and the
vector of features of the first member; and
causing a graphical user interface to display a
recommendation of the job opening to the first member based
on the calculated distance.
REJECTION
Claims 1–20 are rejected under 35 U.S.C. § 101 as being directed to a
judicial exception without significantly more.
ANALYSIS
Patent Eligibility of Claims 1–20
Appellant argues the claims as a group. See Appeal Br. 7–12. We
select claim 1 as representative. See 37 C.F.R. § 41.37(c)(1)(iv).
Regarding claim 1, the Examiner determines each of the limitations of
claim 1 can be performed in the mind as mental processes except for recited
generic components and therefore it recites an abstract idea. Final Act. 3–4.
The Examiner also determines that claim 1 recites additional elements as a
database, computer processor, social networking service, and graphical user
interface at high levels of generality for performing generic functions that do
not integrate claim 1 into a practical application. Id. at 4–5. The Examiner
finds that the processor performs generic functions of storing, computing,
and presenting, and the graphical user interface displays data as insignificant
extra solution activity. Id. at 4–5. The Examiner determines that the recited
databases and social networking service perform insignificant extra solution
activity of data gathering. Id. at 5. Considering these elements individually
or as an ordered combination, the Examiner determines that they apply the
abstract concept to a technical environment in a very general sense that does
not add significantly more to the abstract idea. Id. at 5–6.
Appeal 2020-003327
Application 14/814,397

4
Principles of Law
Section 101 of the Patent Act states:
Whoever 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.
35 U.S.C. § 101. This provision contains an implicit exception: “Laws of
nature, natural phenomena, and abstract ideas are not patentable.” Alice
Corp. v. CLS Bank Int’l, 573 U.S. 208, 216 (2014).
To distinguish patents that claim laws of nature, natural phenomena,
and abstract ideas from those that claim patent-eligible applications, we first
determine whether the claims are directed to a patent-ineligible concept. Id.
at 217. If they are, we consider the elements of each claim, individually and
“as an ordered combination,” to determine if additional elements “‘transform
the nature of the claim’ into a patent-eligible application” as an “inventive
concept” sufficient to ensure the claims in practice amount to significantly
more than a patent on the ineligible concept itself. See id. at 217–18.
The USPTO has issued guidance about this framework. 2019 Revised
Patent Subject Matter Eligibility Guidance, 84 Fed. Reg. 50 (Jan. 7, 2019)
(“Revised Guidance”). Under the Revised Guidance, to determine whether a
claim is “directed to” an abstract idea, we evaluate whether the claim recites:
(1) any judicial exceptions, including certain groupings of abstract ideas
listed in the Revised Guidance (i.e., mathematical concepts, certain methods
of organizing human activities such as a fundamental economic practice, or
mental processes); and (2) additional elements that integrate the judicial
exception into a practical application (see MPEP §§ 2106.05(a)–(c), (e)–(h)
(9th ed. Rev. 10.2019 June 2020) (“MPEP”)). Id. at 52–55.
Appeal 2020-003327
Application 14/814,397

5
Only if a claim (1) recites a judicial exception and also (2) does not
integrate that exception into a practical application, do we then consider
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. Id. at 56.
Step 1: Is Claim 1 Within a Statutory Category?
Claim 1 recites a “method” which is a statutory category of 35 U.S.C.
§ 101, namely, a process. See Final Act. 4. Thus, we consider whether
claim 1 as a whole recites a judicial exception.
Step 2A, Prong One: Does Claim 1 Recite a Judicial Exception?
We determine claim 1 recites certain methods of organizing human
activity by managing relationships or interactions between people or social
activities by following rules and mental processes––concepts performed in
the human mind. Revised Guidance, 84 Fed. Reg. at 52; see October 2019
Update: Subject Matter Eligibility.
Claim 1 recites a method of scoring how closely profiles on a social
networking site match job listings to determine the likelihood of a member
obtaining a job. Spec. ¶¶ 1 (p. 1), 17 (p. 4).2 A machine learning algorithm
assigns weights to profile features of members who obtained a similar job to
create a vector of features for a perfect candidate, and that vector is used to
score how closely member profiles match a perfect candidate. Id. ¶¶ 31–34.

2 Refers to the Specification that was filed July 30, 2015. The Specification
lists duplicate paragraphs 1–19 at pages 1–4 and 5–11. Therefore, citations
to those paragraphs also will cite the page on which the paragraph appears.
Appeal 2020-003327
Application 14/814,397

6
The first step of “using a machine learning algorithm to learn weights,
wherein the learning involves optimizing weights assigned to features to
member profiles, based on sample member profiles of successful candidates,
wherein a successful candidate is a member who previously applied for a
particular job and obtained the particular job” recites a mental process as the
Examiner determines. Appeal Br. 13 (Claims App.); Final Act. 3–4.
This step uses a generic “machine learning algorithm” to replicate the
mental steps that a recruiter or other person would perform mentally to filter
member profiles of candidates who successfully obtained a job to determine
the features that were most relevant to obtaining the job. The Specification
describes how corporate recruiters can search and filter member profiles on
social networks to identify potential candidates who match a particular job
posting and match member profiles to the job postings. Spec. ¶¶ 2 (p. 1), 15
(p. 3). Members search/filter job postings that match their qualifications.
Id. ¶ 15 (p. 3). Recruiters similarly can identify features of candidates who
obtained a job and compare those features to member profiles to determine
how closely a member’s profile matches features of successful candidates as
mental steps. See id.; Revised Guidance, 84 Fed. Reg. at 52 (citing Mayo
Collaborative Servs. v. Prometheus Labs., Inc., 566 U.S. 66, 71 (2012)).
In addition, member profile data organizes a member’s personal and
professional activities, relationships, interactions, and social activities. The
Specification indicates member profiles include personal information such
as name, age, gender, interests, contact information, home town, address,
spouse and family members, educational background (schools, majors), and
employment history with skills, professional organizations, and job titles.
Id. ¶ 9 (p. 7). As such, member profiles also organize human activities.
Appeal 2020-003327
Application 14/814,397

7
Essentially, the profiles of members who obtained jobs are market
survey data to be analyzed and filtered to learn which features led to a job.
Using a generic machine learning algorithm to analyze such market survey
data to identify which features are optimal for a job (see Appeal Br. 7–8) is
an abstract idea. See OIP Techs, Inc. v. Amazon.com, Inc., 788 F.3d 1359,
1361–62 (Fed. Cir. 2015) (claims to analyzing price surveys to identify how
customers responded to offers and estimate an optimal price for a product is
an abstract idea); Revised Guidance, 84 Fed. Reg. at 52 (citing OIP).
The claimed “machine learning algorithm” in described generically.
“[A] machine learning system is utilized that computes a score that provides
a quantitative measurement of the likelihood that a particular member will
obtain a particular job if desired.” Spec. ¶ 17 (p. 4). “The machine learning
system creates virtual candidates/members that are hypothetical ‘perfect fits’
for the jobs in the ecosystem.” Id. ¶ 18 (p. 4).
As described briefly above, machine learning techniques may
be applied to different areas of the system/method described
above to improve reliability. For example, . . . the perfect
candidate generator 316 may receive feedback in the form of
indications of whether or not particular candidates actually
received offers, and then may adjust the one or more policies it
uses to determine the perfect candidate accordingly. For
example, the perfect candidate generator 316 may first indicate
that particular skills A, B, C, and D are key qualities of a
perfect candidate, but then later receive feedback that a certain
number of candidates who applied who had these key qualities
did not receive offers, but candidates, who had an added skill of
E received a high percentage of offers. The perfect candidate
generator 316 may then adjust the policies, not just for this
particular job but potentially for other jobs as well based on this
feedback as to its “performance”.
Id. ¶ 43.
Appeal 2020-003327
Application 14/814,397

8
As recited in claim 1, a machine learning algorithm replicates mental
processes used to identify features of candidates who obtained to a job. No
details of how policies or rules are used to learn weights is claimed. Claims
that merely replicate the mental steps performed by a person, when recited at
such a high level of generality as in claim 1, recite a mental process. See In
re Meyer, 688 F.2d 789, 795 (Fed. Cir. 1982) (holding that claims to a
computerized method of performing a neurological examination by using an
algorithm to represent mental processes that a neurologist should follow to
replace the thinking processes of a neurologist with a computer were not
patent eligible); SmartGene, Inc. v. Advanced Biological Labs., SA, 555 F.
App’x 950, 954–55 (Fed. Cir. 2014) (holding a claimed computer device
replicated thought processes of a medical doctor where it contained, like a
doctor’s mind, “expert rules for evaluating and selecting” from a “plurality
of different therapeutic treatment regimens” and advisory information useful
for the treatment of a patient with different constituents of said different
therapeutic regimens.”).
Appellant argues the claims require a machine learning algorithm to
learn weights and “use of a machine learning algorithm inherently requires
the use of a machine, and cannot be performed in the mind.” Appeal Br. 8.
No details of this process are claimed, however. As claimed, the machine
learning algorithm merely replicates mental processes of recruiters similar to
claims determined to be an abstract idea in Meyer and SmartGene. See Elec.
Power Grp., LLC v. Alstom S.A., 830 F.3d 1350, 1354 (Fed. Cir. 2016) (“In
a similar vein, we have treated analyzing information by steps people go
through in their minds, or by mathematical algorithms, without more, as
essentially mental processes within the abstract-idea category.”).
Appeal 2020-003327
Application 14/814,397

9
The remaining steps except for displaying a recommendation of a job
opening3 also organize human activities in member profiles and analyze the
data by mental process steps. The steps include the following:
receiving, from a jobs database, metadata pertaining to
job listing data corresponding to a job opening, the metadata
including one or more requirements for the job opening,
obtaining, from a member database distinct from the jobs
database, a member profile for a first member of a social
networking service,
obtaining, from the member database, a plurality of
member profiles for members of the social networking service
other than the first member, . . . who have applied for and
obtained job offers for jobs similar to the job opening,
determining, for each of the plurality of member profiles,
a date on which the corresponding member applied for the job
similar to the job opening,
obtaining, for each of the plurality of member profiles, a
version of the corresponding member profile as it was on the
date on which the corresponding member applied for the job
similar to the job opening . . . including a vector of features of
the corresponding member,
applying, . . . one or more of the weights to the vectors
for the plurality of member profiles to obtain a vector of
features representing a perfect candidate for the job opening,
obtaining a vector of features of the first member from
the member profile for the first member, [and]
calculating a distance for the first member with respect to
the perfect candidate for the job opening by comparing the
vector of features.
Appeal Br. 17 (Claims App.).

3 Displaying data is nevertheless extra solution activity and patent ineligible.
See Bilski v. Kappos, 561 U.S. 593, 610–11 (2010) (“Flook stands for the
proposition that the prohibition against patenting abstract ideas ‘cannot be
circumvented by attempting to limit the use of the formula to a particular
technological environment’ or adding ‘insignificant postsolution activity.”’)
(quoting Diamond v. Diehr, 450 U.S. 175, 191–92 (1981)).
Appeal 2020-003327
Application 14/814,397

10
As claimed, such steps of obtaining data such as metadata, member
profiles, and dates recite mental processes and methods of organizing human
activity. Elec. Power, 830 F.3d at 1355 (“But merely selecting information,
by content or source, for collection, analysis, and display does nothing
significant to differentiate a process from ordinary mental processes, whose
implicit exclusion from § 101 undergirds the information-based category of
abstract ideas.”); Synopsys, Inc. v. Mentor Graphics Corp., 839 F.3d 1138,
1146 (Fed. Cir. 2016) (same); see also Content Extraction & Transmission
LLC v. Wells Fargo Bank, Nat’l Ass’n, 776 F.3d 1343, 1347 (Fed. Cir. 2014)
(holding that humans have performed steps of collecting data, recognizing
certain data in the set, and storing recognized data, e.g., by banks reviewing
checks); CyberSource Corp. v. Retail Decisions, Inc., 654 F.3d 1366, 1373
(Fed. Cir. 2011) (“[C]omputational methods which can be performed
entirely in the human mind are the types of methods that embody the ‘basic
tools of scientific and technological work’ that are free to all men and
reserved exclusively to none”)).
In CyberSource, the court held a person could read records of Internet
credit card transactions from a preexisting database, construct a map or list
of transactions made from a particular Internet address, and compare the list
to actual transactions to identify fraudulent transactions that used the same
Internet address to conduct multiple transactions with different credit cards
of different users and different billing addresses. CyberSource, 654 F.3d at
1372–72. Here, claim 1 recites steps a person can perform by writing down
“metadata” of job listings and features of member profiles of successful job
candidates and comparing a “distance” of prospective job seekers based on
the similarity of features in their profiles to features of a perfect candidate.
Appeal 2020-003327
Application 14/814,397

11
The Specification’s description of these features makes clear that they
recite the abstract idea identified above. For example, a “vector” of features
of a member profile merely refers to skills, endorsements or attributes of the
corresponding member. Spec. ¶ 31. Metadata may include requirements for
a job opening. Id. ¶ 45. Job data requirements may be extracted by filters or
thresholds. Id. ¶ 18 (p. 10). Recognizing and extracting data from records is
an abstract idea when recited at such a high level of generality. See Content
Extraction, 776 F.3d at 1347. Filtering data is an abstract idea when recited
at such a high level of generality. See BASCOM Glob. Internet Servs., Inc.
v. AT&T Mobility LLC, 827 F.3d 1341, 1348 (Fed. Cir. 2016) (“We agree
with the district court that filtering content is an abstract idea because it is a
longstanding, well-known method of organizing human behavior, similar to
concepts previously found to be abstract.”); MPEP § 2106.04(a)(2)II.C.
Appellant argues the claims are similar to Example 39 of the 2019
Revised Guidance because they recite the training of a machine learning
algorithm and therefore do not recite a mathematical relationship or a mental
process. Appeal Br. 7–8; Reply Br. 4. Appellant argues that the claims
recite “using a machine learning algorithm to learn weights, wherein the
learning involves optimizing weights assigned to features of member
profiles, based on sample member profiles of successful candidates, wherein
a successful candidate is a member who previously applied for a particular
job and obtained the particular job” and these features are not within one of
the categories of abstract idea in the Revised Guidance. Appeal Br. 8.
As discussed above, using a machine learning algorithm to process
data to learn weights, when recited at a high level of generality as in claim 1
with the other limitations, recites the abstract idea identified above.
Appeal 2020-003327
Application 14/814,397

12
Furthermore, the claim of Example 39 is distinguishable from claim 1
because the steps of “applying one or more transformations to each digital
facial image including mirroring, rotating, smoothing, or contrast reduction
to create a modified set of digital facial images,” creating training sets with
the collected set of digital images, and training the neural network in a first
and a second stage did not recite a mental process because the steps could
not be performed in the human mind. See 2019 Subject Matter Eligibility
Examples: Abstract Ideas, Example 39, at 8–9.
Here, as discussed above, claim 1 recites steps that can be performed
as mental process steps in the human mind or with pen and paper. The steps
also organize the human activity identified above as well. Claim 1 does not
recite steps of training a neural network. The machine learning algorithm is
recited as being used to “learn weights.” Appeal Br. 13 (Claims App.).
Essentially, claim 1 recites a method of creating an index of features
in member profiles of members who obtained a job and using the index to
weight the features (vectors) in an unspecified way. Weighting can result
from features that appear more often in profiles of successful job applicants.
Spec. ¶ 43. Member profiles are collected, parsed, and stored in a search
index to facilitate identification and retrieval of information in response to
search queries. Id. ¶ 15 (p. 9). Unspecified policies (rules) are applied to
create a perfect candidate index used to calculate a distance (score) for a
member profile’s suitability for a particular job. Id. ¶¶ 33, 34, 51.
In an analogous situation, using an index of tags and metafiles to
locate desired information in a database recited longstanding conduct that
existed well before the advent of computer and the Internet. See Intellectual
Ventures I LLC v. Erie Indemnity Co., 850 F.3d 1315, 1327 (Fed. Cir. 2017).
Appeal 2020-003327
Application 14/814,397

13
In Intellectual Ventures, the claims recited a method of organizing and
accessing records by creating an index-searchable database. Id. The court
held that library indexing systems and other classification systems are used
by classifiers to organize and cross-reference resources by certain tags such
as title, author, and subject. Id. Here, claim 1 recites a machine learning
algorithm that classifies (weights) features in member profiles of successful
candidates. Classifiers can perform this step by collecting member profiles,
recording features of candidates who obtained jobs, and weighting features
that appear more frequently in profiles of successful candidates. Spec. ¶ 43.
Just as indexes identify and organize features that appear most often
in a work, the claimed “perfect candidate” indexes features of successful job
candidates. A member profile is compared to a perfect candidate to score
how closely a member profile matches it. See id. at 1328 (“The focus of the
claims, therefore, remains at a high level on searching a database using an
index. The inclusion of XML tags as the chosen index building block, with
little more, does not change that conclusion.”); see also BSG Tech LLC v.
BuySeasons, 899 F.3d 1281, 1286 (Fed. Cir. 2018) (holding the use of
historical usage information to input data into a database was a fundamental,
long-prevalent practice and well-established method of organizing activity);
Berkheimer v. HP Inc., 881 F.3d 1360, 1366–67 (Fed. Cir. 2018) (holding
steps of parsing, comparing, storing, and editing data were not patent
eligible for the reasons set forth in Content Extraction, 776 F.3d at 1347 and
In re TLI Commc’ns LLC Patent Litig. 823 F.3d 607, 613 (Fed. Cir. 2016)).
The claimed “distance” is a concept without parameters. It indicates
how closely a profile matches a perfect candidate in some unclaimed way. It
is adjusted dynamically using policies that are not claimed. Spec. ¶¶ 42–50.
Appeal 2020-003327
Application 14/814,397

14
To predict (score) a prospective member profile’s distance to a job,
the claimed method data mines member profiles of candidates who obtained
the same or similar jobs in the past to create a perfect candidate index that is
used to assess the likelihood of other member profiles obtaining a job listing.
A similar claim to using surveys and crowdsourcing to evaluate a likelihood
of success in a similar marketing context was held to recite an abstract idea
of organizing human activity in Jobin. See In re Jobin, 811 F. App’x 633,
637 (Fed. Cir. 2020). As the court explained in that case:
Despite its expansive language and its recitation of servers and
databases, claim 221 of Jobin’s application is, at bottom,
directed to the collection, organization, grouping, and storage of
data using techniques such as conducting a survey or
crowdsourcing. As the Board correctly concluded, this claim is
directed to a method of organizing human activity—a hallmark
of claims directed to abstract ideas.
Id.
Here, the claimed steps of machine learning weights by an algorithm
and parsing member profiles of successful candidates to identify vectors and
features to be weighted are recited at such a high level of generality that they
recite the abstract idea identified above. Example 39 of the 2019 PEG is
distinguishable because it transformed digital images and created a training
set with the digital images, trained a neural network in a first stage using the
first training set, and created a second training set of images using the first
set of images that were incorrectly detected as facial images (false positive)
and then trained the neural network in a second stage. 2019 PEG, at 8–9.
Accordingly, we determine that claim 1 recites certain methods of
organizing human activity and mental processes identified above and thus
recites a judicial exception under the Revised Guidance.
Appeal 2020-003327
Application 14/814,397

15
Step 2A, Prong Two: Integration into a Practical Application
We next consider whether claim 1 recites any additional elements that
integrate the abstract idea into a practical application. Revised Guidance,
84 Fed. Reg. at 54 (Revised Step 2A, Prong Two). We determine claim 1
lacks additional elements that improve a computer or other technology. The
additional elements do not implement the abstract idea in conjunction with a
particular machine or manufacture that is integral to the claim. They do not
transform or reduce a particular article to a different state or thing. They do
not apply the abstract idea in a meaningful way beyond merely linking it to a
particular technological environment. See Revised Guidance, 84 Fed. Reg.
at 55 and MPEP sections cited therein.
Appellant contends that the Examiner has analyzed only the additional
elements of the database, social networking service, graphical user interface,
and computer processor while ignoring other elements of the claim and thus
has not considered the claim as a whole under Prong Two as required by the
October 2019 Update: Subject Matter Eligibility. Reply Br. 7. Appellant
also asserts that the Examiner has ignored arguments that the present claims
do not monopolize the judicial exception but recite a particular solution to a
particular technical problem described in the Specification using the specific
computer operations in the claims. Id. at 8. Appellant argues that the claims
describe “using a machine learning algorithm to learn weights, wherein the
learning involves optimizing weights based on sample member profiles of
successful candidates” and thus are limited to embodiments where machine
algorithms are used to learn such weights rather than fixed algorithms that
are programmed into a computer by a human as a way to optimize weights
based on member profiles of successful candidates. Id.
Appeal 2020-003327
Application 14/814,397

16
Notably, Appellant does not challenge the Examiner’s findings that
the additional elements are recited as generic components used to perform
generic functions as tools that implement the abstract idea. Final Act. 4;
Ans. 12; Appeal Br. 8–11; Reply Br. 6–8. We agree that claim 1 recites a
database, social networking service, computer processor, and graphical user
interface as generic elements that perform generic functions without any
improvement to computers or other technology claimed. The Examiner has
considered the additional elements in the context of the claim as a whole.
The jobs and member databases store data to be retrieved. Metadata
relating to job listings for a job opening is received from the jobs database.
Member profiles are obtained from the member database. See Appeal Br. 13
(Claims App.). The Specification describes the databases generically. “A
jobs database 302 contains data regarding a plurality of different available
jobs. The data may be defined via one or more job postings collected by a
job posting service.” Spec. ¶ 16 (p. 9). “[A] data layer may include several
databases, such as a database 218 for storing data, including both member
profile data as well as profile data for various organizations (e.g., companies,
schools, etc.).” Id. ¶ 9 (p. 7). “[A] member mapper 310 may select a
corresponding member profile in the member database 312 based on a
matching algorithm that uses various fields (called ‘features’) of metadata
found in the data from the internal database 308.” Id. ¶ 20.
As the Federal Circuit held in BSG, “the recited database structure
similarly provides a generic environment in which the claimed method is
performed.” BSG, 899 F.3d at 1286–87 (“[T]he recitation of a database
structure slightly more detailed than a generic database does not save the
asserted claims at step one.”).
Appeal 2020-003327
Application 14/814,397

17
The social networking service is recited generically. Members belong
to a social networking service. Appeal Br. 13 (Claims App.). Persons and
organizations can register to join a social networking service. Spec. ¶ 9 (p.
7). Other features are described for the social networking service/system in
the Specification but none of the features are claimed. Id. ¶¶ 6–15 (pp. 6–9).
Thus, they cannot integrate the claim into a practical application.
The computer processor is recited generically as applying weights to
vectors for the plurality of member profiles. Appeal Br. 13 (Claims App.).
The Specification indicates processors can be configured by software or as a
hardware-implemented module. Spec. ¶¶ 67–71. A hardware-implemented
module can use a special-purpose processor, general-purpose processor, or a
programmable processor. Id. ¶ 68. However, again, claim 1 does not recite
any feature of the claimed processor beyond the generic function of applying
weights to vectors of member profiles. It is well-settled in this regard that
“mere recitation of a generic computer cannot transform a patent-ineligible
abstract idea into a patent-eligible invention.” Alice, 573 U.S. at 223; see
Bancorp Servs., L.L.C. v. Sun Life Assur. Co. of Can. (U.S.), 687 F.3d 1266,
1277 (Fed. Cir. 2012) (explaining a “computer” is “an automatic electronic
device for performing mathematical or logical operations” and “[a] digital
computer . . . operates on data expressed in digits, solving a problem by
doing arithmetic as a person would do it by head and hand” such that the
“meanings conveniently illustrate the interchangeability of certain mental
processes and basic digital computation and help explain why the use of a
computer in an otherwise patent-ineligible process for no more than its most
basic function––making calculations or computations––fails to circumvent
the prohibition against patenting abstract ideas and mental processes.”).
Appeal 2020-003327
Application 14/814,397

18
The graphical user interface is a generic component used to perform a
generic function of “display a recommendation of the job opening to the first
member based on the calculated distance.” Appeal Br. 14 (Claims App.).
The Specification describes this component as a generic element coupled to
a processor of a mobile device or part of generic computer system 1000 as a
video display, liquid crystal display or cathode ray tube. Spec. ¶¶ 66, 79.
The graphical user interface thus performs insignificant extra solution
activity. It outputs data resulting from the data collection and analysis steps.
Final Act. 5; see SAP Am., Inc. v. InvestPic, LLC, 898 F.3d 1161, 1167 (Fed.
Cir. 2018) (“And ‘merely presenting the results of abstract processes of
collecting and analyzing information, without more (such as identifying a
particular tool for presentation), is abstract as an ancillary part of such
collection and analysis.’ . . . The claims here are directed at abstract ideas
under those principles.”) (citations omitted); see also MPEP § 2106.05(g).
Because claim 1 recites the additional elements generically to perform
generic functions and the Specification confirms that the additional elements
are generic components that do not improve the operation of computers or
other technology, the additional elements do not tie the abstract idea to a
particular machine that is integral to the claim. Instead, they implement the
abstract idea as “tools” without any innovations that integrate the abstract
idea into a practical application. See Ans. 12; Revised Guidance, 84 Fed.
Reg. at 55; MPEP §§ 2106.05(a)I., 2105.05(b)II; buySAFE, Inc. v. Google,
Inc., 765 F.3d 1350, 1354 (Fed. Cir. 2014) (noting that the method claims in
Alice merely invoked “some unspecified, generic computer” to obtain data,
adjust account balances, and issue automated instructions without reciting
any improvement to the functioning of the computer or other technology).
Appeal 2020-003327
Application 14/814,397

19
Appellant also argues that the application solves technical problems in
existing social networking services that connect member profiles without the
ability to determine with accuracy if a member will obtain a job. Appeal Br.
9 (citing Spec. ¶¶ 15, 16). As a result, Appellant asserts that members over
apply for job postings. Spec. ¶ 15 (p. 3). This over subscription wastes time
and resources when members may obtain unnecessary training and education
under the mistaken assumption it will make them competitive for a job when
such training and education actually will not. Id. ¶ 16 (pp. 3–4).
Even if these passages describe a technical problem, e.g., of obtaining
greater accuracy in filtering or matching search results of member profiles to
job postings on social networks, claim 1 does not recite a solution to that
problem. See ChargePoint, 920 F.3d at 769–70 (“Even if ChargePoint’s
specification had provided, for example, a technical explanation of how to
enable communication over a network for device interaction (which, as
discussed above, it did not), the claim language here would not require those
details. Instead, the broad claim language would cover any mechanism for
implementing network communication on a charging station.”); Ericsson
Inc. v. TCL Commc’ns Tech. Holdings Ltd., 955 F.3d 1317, 1325 (Fed. Cir.
2020) (“[T]he specification may be helpful in illuminating what a claim is
directed to [but it] must always yield to the claim language when identifying
the ‘true focus of a claim.’”) (citation omitted); Two-Way Media Ltd. v.
Comcast Cable Commc’ns, LLC, 874 F.3d 1329, 1339 (Fed. Cir. 2017)
(rejecting argument the claim solved technical problems of excessive loads,
network congestion, variations in delivery times, scalability, and imprecise
record keeping where the claim used only generic functional language to
achieve these purported solutions).
Appeal 2020-003327
Application 14/814,397

20
The machine learning algorithm is claimed generically. It learns and
optimizes weights. Appeal Br. 13 (Claims App.). These steps are recited at
such a high level of generality without any innovative features that they are
part of the abstract idea identified under Prong One. Even if an innovation is
described in the Specification, claim 1 does not recite any innovation. Thus,
“the complexity of the implementing software or the level of detail in the
specification does not transform a claim reciting only an abstract concept
into a patent-eligible system or method.” Accenture Global Servs. GmbH v.
Guidewire Software, Inc., 728 F.3d 1336, 1345 (Fed. Cir. 2013).
Reciting concepts at this level of generality recites a judicial exception
rather than an integration into a practical application. See OIP, 788 F.3d at
1362–63 (holding that claims to a method of offer-based price optimization
recited an abstract idea and limiting the claims to price optimization without
preempting all price optimization did not make the claims any less abstract).
In Gopalan, the court determined that optimization of parameters to
yield optimal true positives and false positives recited an abstract concept.
The claims only generically recite “a metric,” “an optimization
technique,” an “optimization parameter,” “a value of a number
of independent measures,” and “a value for a confidence
measure.” None of these variables are defined, and the claims do
not concretely limit these variables such that the claims do not
merely claim the result of obtaining a “substantially optimal
combination of true positives and false positives” in the data set.
. . .
Thus, the claims do not “embody a concrete solution to a
problem” because they lack “the specificity required to
transform a claim from one claiming only a result to one
claiming a way of achieving it.”
In re Gopalan, 809 F. App’x 942, 943–44, 946 (Fed. Cir. 2020).
Appeal 2020-003327
Application 14/814,397

21
Example 39 of the 2019 PEG illustrates why the machine learning
algorithm recites a judicial exception rather than a technical improvement
sufficient to integrate the judicial exception into a practical application. In
Example 39, the claimed method collects a set of digital facial images from
a database, transforms each digital facial image by using one or more of the
specifically claimed techniques of rotating, smoothing, or contrast reduction
to create a modified set of digital facial images, creates a first training set of
the collected set of digital facial images, the modified set of digital facial
images, and a set of digital non-facial images, and trains the neural network
in a first stage. Subject Matter Eligibility Examples: Abstract Idea, at 8.
Then, a second training set is created for a second stage of training using the
first training set and the digital non-facial images that the neural network
incorrectly detected as facial images after the first training stage. Id. at 9.
Here, the machine learning algorithm learns weights of features in
some unclaimed manner. No training set is used to teach the machine
learning algorithm. Thus, it is not clear how the learning provides more
accurate results when no structured learning process or data set is used to
provide a particular level of training. Learning weights in an unspecified
manner does not ensure accuracy when the Specification describes features
as including a wide variety of data in a member profile. Nor does claim 1
recite how the weights are used to obtain vectors of features for a “perfect
candidate” to ensure that accuracy is provided in the distance calculation for
a member profile. Member profiles are aggregated by applying a policy and
weights learned from a machine learning algorithm to create a hypothetical
perfect candidate” but no technical details are claimed. Spec. ¶ 32. Some
feedback may be received but such details are not claimed. See id. ¶ 43.
Appeal 2020-003327
Application 14/814,397

22
If the requirements of Section 101 could be circumvented simply by
appending a generic “machine learning algorithm” or generic “processor” to
a claim that recites an abstract idea, the § 101 analysis would become an
exercise in drafting.
The fact that a computer “necessarily exist[s] in the physical,
rather than purely conceptual, realm,” Brief for Petitioner 39, is
beside the point. There is no dispute that a computer is a
tangible system (in § 101 terms, a “machine”), or that many
computer-implemented claims are formally addressed to patent-
eligible subject matter. But if that were the end of the § 101
inquiry, an applicant could claim any principle of the physical
or social sciences by reciting a computer system configured to
implement the relevant concept. Such a result would make the
determination of patent eligibility “depend simply on the
draftsman’s art,” . . . thereby eviscerating the rule that “‘[l]aws
of nature, natural phenomena, and abstract ideas are not
patentable.’”
Alice, 573 U.S. at 224.
Accordingly, we determine that claim 1 lacks any additional elements
that are sufficient to integrate the abstract idea into a practical application.
Step 2B: Does Claim 1 Include an Inventive Concept?
We next consider if claim 1 recites additional elements, individually,
or as an ordered combination, that provide an inventive concept. Alice, 573
U.S. at 217–18. This step is satisfied when claim limitations involve more
than well-understood, routine, and conventional activities previously known
in industry. Berkheimer v. HP Inc., 881 F.3d 1360, 1367 (Fed. Cir. 2018);
see Revised Guidance, 84 Fed. Reg. 56 (explaining that the second step of
the Alice analysis considers whether a claim adds a specific limitation that is
beyond the recited judicial exception and that also is not “well-understood,
routine, conventional” activity in the field).
Appeal 2020-003327
Application 14/814,397

23
Considered individually, the limitations of claim 1 recite the abstract
idea identified above. The Specification describes the additional elements as
generic components that perform well-understood, routine, and conventional
functions. It provides no indication that Appellant invented or improved
these computer components but instead uses them to perform their known
functions as tools to implement the abstract idea identified in Prong One.
As an ordered combination, these elements provide no more than
when they are considered individually. Alice, 573 U.S. at 225. They are
used as tools to implement the judicial exception. See SAP, 898 F.3d at
1169–70 (claimed databases and processors did not improve computers but
used available computers and functions as tools to execute the claimed
process); Inventor Holdings, LLC v. Bed Bath & Beyond, Inc., 876 F.3d
1372, 1378 (Fed. Cir. 2017) (considering the steps of representative claims
as an “ordered combination” reveals they “amount to ‘nothing significantly
more’ than an instruction to apply [an] abstract idea” using generic computer
technology).
Even if the steps are groundbreaking, innovative, or brilliant, that is
not enough for eligibility. See Ass’n for Molecular Pathology v. Myriad
Genetics, Inc., 569 U.S. 576, 591 (2013); accord SAP Am., 898 F.3d at 1163
(“No matter how much of an advance in the finance field the claims recite,
the advance lies entirely in the realm of abstract ideas, with no plausibly
alleged innovation in the non-abstract application realm. An advance of that
nature is ineligible for patenting.”); Ultramercial, Inc. v. Hulu, LLC, 772
F.3d 709, 716 (Fed. Cir. 2014) (where all eleven steps instructed artisan to
implement abstract idea with routine, conventional activities, that some steps
were not used previously in this art did not confer patent eligibility).
Appeal 2020-003327
Application 14/814,397

24
“It is well-settled that placing an abstract idea in the context of a
computer does not ‘improve’ the computer or convert the idea into a patent-
eligible application of that idea.” Interval Licensing LLC v. AOL, Inc., 896
F.3d 1335, 1346 (Fed. Cir. 2018). A similar claim to an online method of
matching loan applications of borrowers with lenders was not patent eligible.
We find that claim 1 does not recite any elements that
individually, or as an ordered combination, transform the
abstract idea of coordinating loans into a patent-eligible
application of that idea. “At best, the claim[ ] describe[s] the
automation of [a] fundamental economic concept . . . through
the use of generic-computer functions.” OIP Techs., [Inc. v.
Amazon.com, Inc., 788 F.3d 1359, 1363 (Fed. Cir. 2015)]. It is
well settled, though, that automating conventional activities
using generic technology does not amount to an inventive
concept. See Alice, 134 S. Ct. at 2358 (explaining that “if a
patent’s recitation of a computer amounts to a mere instruction
to implement an abstract idea on . . . a computer, that addition
cannot impart patent eligibility”) . . .; Intellectual Ventures, 792
F.3d at 1367 (“claiming the improved speed or efficiency
inherent with applying the abstract idea on a computer [does
not] provide a sufficient inventive concept”); Bancorp Servs.,
L.L.C. v. Sun Life Assur. 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.”).
LendingTree, 656 F. App’x at 991; see also Jobin, 811 F. App’x. at 637–38
(“[T]he ‘online system,’ ‘server,’ ‘data structure,’ and ‘user device’ elements
recite generic technology for implementing the claimed abstract idea. The
Board correctly concluded that, considered individually or as an ordered
combination, the additional elements in Jobin’s claim 221 do not transform
the claim into a patent eligible application of the abstract idea.”).

Appeal 2020-003327
Application 14/814,397

25
Even if the claims do not monopolize the abstract idea as Appellant
asserts (Reply Br. 8), the Examiner’s § 101 analysis addresses Appellant’s
argument. “[W]hile preemption may signal patent ineligible subject matter,
the absence of complete preemption does not demonstrate patent eligibility.”
FairWarning IP, LLC v. Iatric Sys., Inc., 839 F.3d 1089, 1098 (Fed. Cir.
2016) (quoting Ariosa Diagnostics, Inc. v. Sequenom, Inc., 788 F.3d 1371,
1379 (Fed. Cir. 2015)); OIP, 788 F.3d at 1362–63 (“[T]hat the claims do not
preempt all price optimization or may be limited to price optimization in the
e-commerce setting do not make them any less abstract.”). Furthermore, as
the court held in Ariosa, “[w]here 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.”
Ariosa, 788 F.3d at 1379.
Accordingly, we determine that claim 1 lacks an inventive concept
sufficient to transform the abstract idea into patent eligible subject matter.
Thus, we sustain the rejection of claims 1–20.
CONCLUSION
In summary:
Claims
Rejected
35 U.S.C.
§
Reference(s)/
Basis
Affirmed Reversed
1–20 101 Eligibility 1–20
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a). See 37 C.F.R.
§ 1.136(a)(1)(iv).
AFFIRMED