Google Inc.

Patent Trials and Appeals Board

Jun 26, 2020

2019002937 (P.T.A.B. Jun. 26, 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.
14/230,222 03/31/2014 Yifang Liu 16113-5363001 5793
26192 7590 06/26/2020
FISH & RICHARDSON P.C.
PO BOX 1022
MINNEAPOLIS, MN 55440-1022
EXAMINER
FIGUEROA, KEVIN W
ART UNIT PAPER NUMBER
2124
NOTIFICATION DATE DELIVERY MODE
06/26/2020 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):
PATDOCTC@fr.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte YIFANG LIU
Appeal 2019-002937
Application 14/230,222
Technology Center 2100
Before ALLEN R. MACDONALD, JEREMY J. CURCURI, and
PHILLIP A. BENNETT, Administrative Patent Judges.
CURCURI, Administrative Patent Judge.
DECISION ON APPEAL
STATEMENT OF THE CASE
Pursuant to 35 U.S.C. § 134(a), Appellant1 appeals from the
Examiner’s decision to reject claims 1 and 28–54. We have jurisdiction
under 35 U.S.C. § 6(b).
We REVERSE.

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 Google LLC.
Appeal Br. 1.
Appeal 2019-002937
Application 14/230,222
2
CLAIMED SUBJECT MATTER
The claims are directed to “determining mutual information with
absolute dependency for feature selection.” Abstract. Claim 1, reproduced
below, is illustrative of the claimed subject matter:
1. A computer-implemented method performed by
one or more computers of a machine learning system, the
method comprising:
receiving, by the one or more computers of the machine
learning system, a dataset comprising multiple items of a same
type, wherein each item has multiple properties;
determining, for each of multiple iterations, a value
representing mutual information between a first property and a
second property, wherein, for each iteration, determining the
value comprises:
selecting, by the one or more computers of the
machine learning system, one of the multiple properties
as the first property and another of the multiple properties
as the second property; and
determining, by the one or more computers of the
machine learning system, the value representing the
mutual information between the first property and the
second property that are selected for the iteration,
wherein the value representing the mutual information is
generated using a summation of terms that are generated
using logarithm values, and wherein the values of the
terms are generated using an absolute value operation
that disallows negative logarithm values from decreasing
the summation;
based on the determined values representing mutual
information between the selected properties from the multiple
properties, selecting, by the one or more computers of the
machine learning system, a subset of the multiple properties for
use in the machine learning system;
Appeal 2019-002937
Application 14/230,222
3
receiving, by the one or more computers of the machine
learning system, another item of the same type as the multiple
items from the dataset, wherein the other item has a known
value for at least one of the subset of properties;
predicting, by the one or more computers of the machine
learning system, a value for a property of the other item having
an unknown value, wherein the value is predicted using a
machine learning model of the machine learning system based
on the known value for the at least one of the subset of
properties; and
providing, by the one or more computers of the machine
learning system, the predicted value over a network to a client
device for display on a user interface of the client device.
REJECTIONS
Claims 1 and 28–54 stand rejected under 35 U.S.C. § 101 as directed
to a judicial exception without significantly more. Final Act. 2–5.
OPINION
Claims 1, 44, and 53 are the independent claims. The Examiner
determines the following:
Claim(s) 1, 44, and 53 is/are directed to receiving a dataset, the
dataset having multiple items having multiple properties,
determining a value representing mutual information between
the properties for each iteration, receiving an item of the same
type, predicting a value for an unknown property, and providing
the value on display. This is an abstract idea of receiving data,
applying mathematical manipulations to it, predicting a value for
an unknown property, and displaying that value. This is similar
to known abstract ideas as “Collecting information, analyzing it,
and displaying certain results of the collection and analysis
(Electric Power Group)”. For example, the claims receive a data
set, i.e. collecting information, the claims determine the mutual
information between the values of the dataset and from that,
Appeal 2019-002937
Application 14/230,222
4
predict a value based on the known value, i.e. analyzing it, and
providing a display, i.e. displaying result of the analysis.
Final Act. 2; see also Final Act. 3–5 (determining claims 1 and 28–54 are
not directed to significantly more than the abstract idea), Ans. 3–7.
An invention is patent-eligible if it claims a “new and useful process,
machine, manufacture, or composition of matter.” 35 U.S.C. § 101.
However, the U.S. 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. E.g., Alice Corp. v. CLS Bank Int’l, 573
U.S. 208, 216 (2014).
In determining whether a claim falls within an excluded category, we
are guided by the Court’s two-part framework, described in Mayo and Alice.
Id. at 217–18 (citing Mayo Collaborative Servs. v. Prometheus Labs., Inc.,
566 U.S. 66, 75–77 (2012)). In accordance with that framework, we first
determine what concept the claim is “directed to.” See Alice, 573 U.S. at 219
(“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
determined to be patent eligible include physical and chemical processes,
such as “molding rubber products” (Diamond v. Diehr, 450 U.S. 175, 191
Appeal 2019-002937
Application 14/230,222
5
(1981)); “tanning, dyeing, making waterproof 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
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 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. (citation omitted)
(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 (quotation marks omitted).
“A claim that recites an abstract idea must include ‘additional features’ to
ensure ‘that the [claim] is more than a drafting effort designed to
monopolize the [abstract idea].’” Id. (alterations in original) (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.
Appeal 2019-002937
Application 14/230,222
6
In January 2019, the U.S. Patent and Trademark Office (USPTO)
published revised guidance on the application of § 101. 2019 Revised Patent
Subject Matter Eligibility Guidance, 84 Fed. Reg. 50 (Jan. 7, 2019)
(“2019 Revised Guidance”).2 “All USPTO personnel are, as a matter of
internal agency management, expected to follow the guidance.” Id. at 51; see
also October 2019 Update at 1.
Under the 2019 Revised Guidance and the October 2019 Update, 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) (“Step 2A, Prong One”); and
(2) additional elements that integrate the judicial exception into a
practical application (see MPEP § 2106.05(a)–(c), (e)–(h) (9th ed.
Rev. 08.2017, Jan. 2018)) (“Step 2A, Prong Two”).3
2019 Revised Guidance, 84 Fed. Reg. at 52–55.

2 In response to received public comments, the Office issued further
guidance on October 17, 2019, clarifying the 2019 Revised Guidance.
USPTO, October 2019 Update: Subject Matter Eligibility (the “October
2019 Update”) (available at
https://www.uspto.gov/sites/default/files/documents/peg_oct_2019_update.p
df).
3 This evaluation is performed by (a) identifying whether there are any
additional elements recited in the claim beyond the judicial exception, and
(b) evaluating those additional elements individually and in combination to
determine whether the claim as a whole integrates the exception into a
practical application. See 2019 Revised Guidance - Section III(A)(2), 84
Fed. Reg. 54–55.
Appeal 2019-002937
Application 14/230,222
7
Only if a claim (1) recites a judicial exception and (2) does not
integrate that exception into a practical application, do we then look, under
Step 2B, 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.
2019 Revised Guidance, 84 Fed. Reg. at 52–56.
Are the claims patent-eligible?
Step 1
Claim 1 recites a method, which falls with the “process” category of
35 U.S.C. § 101. Claim 44 recites a system, which falls within the
“machine” category of 35 U.S.C. § 101. Claim 53 recites a computer-
readable medium, which falls within the “manufacture” category of 35
U.S.C. § 101. Thus, we must determine whether the claim recites a judicial
exception and fails to integrate the exception into a practical application. See
2019 Revised Guidance, 84 Fed. Reg. at 54–55. If both elements are
satisfied, the claim is directed to a judicial exception under the first step of
the Alice/Mayo test. See id.
Step 2A, Prong One
Claim 1 is a method claim, which recites, in part, the following:
determining, by the one or more computers of the machine
learning system, the value representing the mutual information
between the first property and the second property that are
Appeal 2019-002937
Application 14/230,222
8
selected for the iteration, wherein the value representing the
mutual information is generated using a summation of terms that
are generated using logarithm values, and wherein the values of
the terms are generated using an absolute value operation that
disallows negative logarithm values from decreasing the
summation.
This step covers determining mutual information4 of two random
variables, and is thus one or more of “mathematical relationships,
mathematical formulas or equations, mathematical calculations.” 2019
Revised Guidance, 84 Fed. Reg. 52. These recitations are thus the abstract
idea of “[m]athematical concepts.” Id.
Claims 44 and 53 recite limitations similar to those discussed with
respect to claim 1. Thus, claims 44 and 53 also recite an abstract idea.
Step 2A—Prong Two
Because claims 1, 44, and 53 recite a judicial exception, we next
determine if the claims recite additional elements that integrate the judicial
exception into a practical application.
In the “Background” section, Appellant’s Specification discloses
known techniques for calculating mutual information of random variables.
Spec. ¶¶ 3–4.
In the “Background” section, Appellant’s Specification further
discloses a first case of two random variables where “the mutual information
I(X,Y)=0.15320219422. Using a Naïve Bayes classifier, the simple
classification accuracy may be 0.910.” Spec. ¶ 6
In the “Background” section, Appellant’s Specification further
discloses a second case of two random variables where “the mutual

4 The claim recites a modified mutual information calculation that uses an
absolute value operation.
Appeal 2019-002937
Application 14/230,222
9
information I(X,Y)=0.15666606782. Using a Naïve Bayes classifier, the
simple classification accuracy may be 0.908.” Spec. ¶ 7.
In the “Background” section, Appellant’s Specification discloses the
following need:
The first case may present a greater relation between X
and Y than the second case, as the simple classification accuracy
of the first case is 91 %, while in the second case it’s 90.8%.
However, the mutual information values determined according
to (1) provide a contradictory result, with the second case having
more mutual information than the first case. The positive and
negative values of the log term may cancel out in the first case,
due to the slope of the log function around y = 0, resulting in the
value for mutual information for the first case being lower than
it should be.
Spec. ¶ 8.
Appellant’s Specification discloses the following solution: a modified
mutual information calculation that uses an absolute value operation. Spec. ¶
9. Appellant’s Specification further discloses “[t]he first of the two random
variables may be a feature of [an] item, and the second of the two random
variables may be a label for the item. The [modified] mutual information
may be used in a machine learning system to predict a value for at least one
of the random variables for an item for which the value is unknown.” Spec. ¶
10. Appellant’s Specification further discloses “[s]ystems and techniques
disclosed herein may allow for the determination of [modified] mutual
information with absolute dependency for feature selection in machine
learning models.” Spec. ¶ 16.
With respect to applying the modified mutual information calculation
that uses an absolute value operation to feature selection in a machine
Appeal 2019-002937
Application 14/230,222
10
learning system that predicts a value and provides the predicted value to a
user, claim 1 recites the following:
based on the determined values representing mutual
information between the selected properties from the multiple
properties, selecting, by the one or more computers of the
machine learning system, a subset of the multiple properties for
use in the machine learning system;
receiving, by the one or more computers of the machine
learning system, another item of the same type as the multiple
items from the dataset, wherein the other item has a known
value for at least one of the subset of properties;
predicting, by the one or more computers of the machine
learning system, a value for a property of the other item having
an unknown value, wherein the value is predicted using a
machine learning model of the machine learning system based
on the known value for the at least one of the subset of
properties; and
providing, by the one or more computers of the machine
learning system, the predicted value over a network to a client
device for display on a user interface of the client device.
We conclude these limitations integrate the recited judicial exception
of a mathematical concept into a practical application. Under the guidance, a
judicial exception may be integrated into a practical application where it
provides “an improvement to . . . any other technology or technical field.”
MPEP § 2106.05(a). Here, these additional limitations provide an
improvement to the technical field of machine learning systems by
improving feature selection to improve prediction performance. As
explained in the Specification, “[t]he determination of mutual information
with absolute dependency between two random variables may be optimized
for accuracy, increasing prediction precision.” Spec. ¶ 29. Thus, claim 1 is
not directed to an abstract idea.
Appeal 2019-002937
Application 14/230,222
11
Claims 44 and 53 recite limitations similar to those discussed with
respect to claim 1. Thus, claims 44 and 53 are also not directed to an abstract
idea.
Thus, claims 1, 44, and 53 recite additional elements that integrate the
judicial exception into a practical application.
Accordingly, we are persuaded by Appellant’s argument that:
In the present application, the claims provide clear
improvements to the way computer-based machine learning
models are built and operate. In claim 1, the improvements
include determining a “value representing the mutual
information between the first property and the second property
... generated using a summation of terms that are generated using
logarithm values ... using an absolute value operation that
disallows negative logarithm values from decreasing the
summation,” selecting “a subset of the multiple properties for use
in the machine learning system” “based on the determined values
representing mutual information between the selected properties
from the multiple properties,” and predicting “a value for a
property ... having an unknown value ... using a machine learning
model of the machine learning system based on the known value
for the at least one of the subset of properties.” These features,
among others, improve machine learning systems to be able to
provide “better feature selection” and consequently better
predictions than the prior systems, as the Specification discusses.
Appeal Br. 10 (citing Spec. ¶ 23); see also Reply Br. 1–2 (“[T]he claims
recite a specific technique showing how the machine learning system can
better select the properties or features to be used in a machine learning
model.”).
We, therefore, do not sustain the Examiner’s rejection of claim 1.
We also do not sustain the Examiner’s rejection of claims 28-43 and
54, which depend from claim 1.
Appeal 2019-002937
Application 14/230,222
12
Claim 44 recites similar limitations as claim 1. We also do not sustain
the Examiner’s rejection of claim 44 for the same reasons discussed above
when addressing claim 1.
We also do not sustain the Examiner’s rejection of claims 45–52,
which depend from claim 44.
Claim 53 recites similar limitations as claim 1. We also do not sustain
the Examiner’s rejection of claim 53 for the same reasons discussed above
when addressing claim 1.
CONCLUSION
The Examiner’s rejection is reversed.
DECISION SUMMARY
In summary:
Claims
Rejected
35 U.S.C.
§
Basis Affirmed Reversed
1, 28–54 101 Judicial Exception 1, 28–54
REVERSED