Ex Parte Satish et al

Patent Trial and Appeal Board

Sep 12, 2017

13742218 (P.T.A.B. Sep. 12, 2017)

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.
13/742,218 01/15/2013 Sourabh Satish 32745-21981/US 8354
34415 7590 09/14/2017
SYMANTRfV FF.NWTPK
EXAMINER
SILICON VALLEY CENTER PEREZ-ARROYO, RAQUEL
801 CALIFORNIA STREET
MOUNTAIN VIEW, CA 94041 ART UNIT PAPER NUMBER
2169
NOTIFICATION DATE DELIVERY MODE
09/14/2017 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):
ptoc @ fenwick.com
bhoffman @ fenwick.com
sfuentes @ fenwick. com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte SOURABH SATISH, GOVIND SALINAS, and
VINCENT CHEONG
Appeal 2017-005726
Application 13/742,2181
Technology Center 2100
Before JEFFREY S. SMITH, JOHNNY A. KUMAR, and
TERRENCE W. McMILLIN, Administrative Patent Judges.
McMILLIN, Administrative Patent Judge.
DECISION ON APPEAL
This is a decision on appeal under 35 U.S.C. § 134(a) of the final
rejection of claims 1—17, 19, and 20. Final Act. 1. We have jurisdiction
under 35 U.S.C. § 6(b).
We AFFIRM.
1 According to Appellants, the real party in interest is Veritas Technologies
LLC (App. Br. 2).
Appeal 2017-005726
Application 13/742,218
THE CLAIMED INVENTION
The present invention generally relates to computer security and more
particularly to classifying samples, such as samples of computer files. Spec.
11. Independent claim 1 is directed to computer-implemented method;
independent claim 9 is directed to a computer; and independent claim 15 is
directed to a non-transitory computer-readable storage medium. App. Br.
12, 14, 16.
Claim 1 recites:
1. A computer-implemented method of classifying a sample,
comprising:
establishing a set of samples containing labeled and
unlabeled samples;
gathering values of features from the labeled and
unlabeled samples;
selecting a subset of the features;
clustering together labeled and unlabeled samples having
at least a threshold measure of similarity among the gathered
values of the selected subset of features to produce a set of
clusters, each cluster having a subset of samples from the set of
samples;
recursively iterating the selecting and clustering steps on
the subset of samples in each cluster in the set of clusters until at
least one stopping condition is reached, wherein there are a
plurality of iterations of the selecting and clustering steps, a first
clustering iteration uses a first threshold measure of similarity
and subsequent clustering iterations use increasingly stricter
similarity thresholds, the iterations produce a cluster having a
labeled sample and an unlabeled sample, and the selecting step
selects different subsets of the features for different iterations of
the plurality of iterations by:
2
Appeal 2017-005726
Application 13/742,218
determining an amount of variance in values of
features of samples in a cluster of the set of clusters
produced by a clustering iteration; and
selecting a subset of features for a subsequent
clustering iteration from among available features
responsive to the determined amount of variance in values
of features of samples in the cluster of the set of clusters
produced by the clustering iteration; and
propagating a label from the labeled sample in the cluster
to the unlabeled sample in the cluster to classify the unlabeled
sample.
REJECTIONS ON APPEAL
Claims 1, 2, 4, 8—11, 14—17, and 20 stand rejected under 35 U.S.C.
§ 103(a) as being unpatentable over Ribnick et al. (US 2013/0202200 Al,
published Aug. 8, 2013) (“Ribnick”), Gutta et al. (US 6,801,917 B2,
published Oct. 5, 2004) (“Gutta”), and Ben (US 8,473,532 Bl, published
June 25, 2013). Final Act. 2.
Claims 3, 5, 6, 12, 13, and 19 stand rejected under 35 U.S.C. § 103(a)
as being unpatentable over Ribnick, Gutta, and Schipka (US 2009/0013405
Al, published Jan. 8, 2009).2 Final Act. 23.
Claim 7 stands rejected under 35 U.S.C. § 103(a) as being
unpatentable over Ribnick, Gutta, and Han et al. (US 8,515,193 Bl,
published Aug. 20, 2013).3 Final Act. 33.
2 The statement of rejection lists Ribnick, Gutta, and Schipka, but should
properly also list Ben, which is relied upon in the grounds of rejection for
independent claims 1, 8, and 15, upon which these claims depend. See Final
Act. 23. We find the Examiner’s omission of Ben to be harmless error.
3 The statement of rejection lists Ribnick, Gutta, and Han, but should
properly also list Ben, which is relied upon in the grounds of rejection for
3
Appeal 2017-005726
Application 13/742,218
ANALYSIS
We have reviewed the Examiner’s rejections in light of Appellants'
arguments that the Examiner erred. We are not persuaded that Appellants
identify reversible error. Upon consideration of the arguments presented in
the Appeal Brief, we agree with the Examiner that all the pending claims are
unpatentable over the cited combination of references. We adopt as our own
the findings and reasons set forth in the rejection from which this appeal is
taken and in the Examiner’s Answer. We provide the following explanation
to highlight and address specific arguments and findings primarily lor
emphasis.
Regarding claim 1, Appellants contend “Ben does not teach or suggest
that ‘a first clustering iteration uses a first threshold measure of similarity
and subsequent clustering iterations use increasingly stricter similarity
threshold’ as claimed.” App. Br. 8. Specifically, Appellants argue Ben
teaches “that the threshold is ‘dynamically adjusted]... such that the process
will produce a group that is within the desired size’” which “may include
decreasing the threshold (i.e., using less strict similarity thresholds).” Id.
As cited by the Examiner, Ben discloses:
The process starts by identifying the total number of items “n ”
(files or folders) to be clustered. It compared each pair of items
to determine how similar the pairs of items are.... Based on the
results, the process then determines a threshold in any
conventional manner such as by taking the average similarity,
the median or another percentile as the threshold. The process
clusters files by partitioning the n given items into smaller and
smaller groups. It starts with the n items as the initial group. It
independent claims 1, 8, and 15, upon which these claims depend. See Final
Act. 33. We find the Examiner’s omission of Ben to be harmless error.
4
Appeal 2017-005726
Application 13/742,218
compares each pair of items in the group. If a pair (j, k) of items
in the group have a similarity less than the threshold, then it
splits the group into two groups. . . . The process is continued in
an iterative manner until a group is found wherein all pairs of
items have a similarity larger than the threshold. . . . The limit
is use to dynamically adjust the threshold such that the process
will produce a group that is within the desired size. As shown in
the pseudo code, the threshold is increased if all resulting groups
are too large and is decreased if all resulting groups are too
small.
Ben col. 8,11. 47—62, 65—67, col. 9,11. 8—12 (emphasis added).
The Examiner finds, and we agree, Ben teaches “iterative clustering
based on feature similarity threshold with first threshold value” and
“dynamically increasing similarity threshold to be used in the clustering
iterations, where increasing the similarity threshold means that the similarity
amongst the clustered items must be even greater, in other words, stricter.”
Final Act. 7.
Appellants contend relying on Ben’s teaching of a clustering criteria
including producing a group that is within a desired size is based on
impermissible hindsight. App. Br. 9. Specifically, Appellants argue that
Ben’s teaching to “decrease the similarity threshold in certain
circumstances” would lead away from the claimed invention, and Ben’s
clustering iterations using increasingly strict thresholds is not sufficient to
establish inherency. App. Br. 10.
We note a reference does not teach away if it merely expresses a
general preference for an alternative invention but does not “criticize,
discredit, or otherwise discourage” investigation into the invention claimed.
In re Fulton, 391 F.3d 1195, 1201 (Fed. Cir. 2004). Ben does teach both
that the “threshold is increased if all resulting groups are too large” and that
5
Appeal 2017-005726
Application 13/742,218
the threshold “is decreased if all resulting groups are too small” in order to
“produce a group that is within the desired size.” Ben col. 9,11. 9-12.
However, Ben’s teaching to decrease the threshold if all resulting groups are
too small does not criticize, discredit, or otherwise discourage increasing the
threshold if all resulting groups are too large. Ben merely provides an
alternative to increasing the threshold. We agree with the Examiner’s
findings that the “fact that Ben is performing an additional step of evaluating
the minimum limit, does not preclude that [] Ben’s threshold is based on
similarity” and “that it can be adjusted to a higher or increased threshold.”
Ans. 2.
Appellants’ contend the Examiner’s rejection relies upon
impermissible hindsight, because the Examiner’s proffered rationale for the
combination, specifically that Ben’s clustering produces “a group that is
within the desired size,” does not address the Appellants’ motivation for the
claimed invention of “‘caus[ing] only relatively similar file samples to be
clustered together and result in more clusters each containing fewer samples’
in subsequent clustering iterations.” App. Br. 9 (citing Spec. 1 54). Whether
Appellants may have been driven by a different motivation than the rationale
supplied by the Examiner does not negate obviousness. See KSR Int 7 Co. v.
Teleflex Inc., 550 U.S. 398, 419 (2007) (“In determining whether the subject
matter of a patent claim is obvious, neither the particular motivation nor the
avowed purpose of the patentee controls. What matters is the objective
reach of the claim.”).
Therefore, we find Appellants’ allegation of impermissible hindsight
is unsupported in the record. Appellants provide no evidence that
combining the teachings of Ribnick, Gutta, and Ben, as proffered by the
6
Appeal 2017-005726
Application 13/742,218
Examiner (Final Act. 7), would have been “uniquely challenging or difficult
for one of ordinary skill in the art” {Leapfrog Enters. Inc. v. Fisher-Price,
Inc., 485 F.3d 1157 (Fed. Cir. 2007)), nor have Appellants provided any
objective evidence of secondary considerations, which our reviewing court
guides “operates as a beneficial check on hindsight.” Cheese Systems, Inc.
v. Tetra Pak Cheese and Powder Systems, Inc., 725 F.3d 1341, 1352 (Fed.
Cir. 2013). We are further unpersuaded of Examiner error regarding Ben’s
teachings being insufficient to establish inherency because the Examiner did
not make inherency findings.
Appellants have not provided persuasive evidence that clustering
items starting with an initial measure of similarity for an initial group, and
iteratively clustering iterations requiring more similarity and resulting in
smaller groups, as required by claim 1, is not taught or otherwise suggested
by Ben’s clustering n items with an initial group of items and partitioning
the items into smaller and smaller groups utilizing similarity thresholds.
Accordingly, we sustain the § 103 rejection of claim 1, as well as the
rejection of claims 2—17, 19, and 20 not separately argued.
DECISION
The rejections of claims 1—17, 19, and 20 are affirmed.
No time period for taking any subsequent action in connection with this
appeal may be extended under 37 C.F.R. § 1.136(a)(l)(iv).
AFFIRMED
7