Ex Parte Deolalikar et al

Patent Trial and Appeal Board

May 4, 2016

13383592 (P.T.A.B. May. 4, 2016)

UNITED STA TES p A TENT AND TRADEMARK OFFICE
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR
13/383,592 01112/2012 Vinay Deolalikar
56436 7590 05/06/2016
Hewlett Packard Enterprise
3404 E. Harmony Road
Mail Stop 79
Fort Collins, CO 80528
UNITED STATES DEPARTMENT OF COMMERCE
United States Patent and Trademark Office
Address: COMMISSIONER FOR PATENTS
P.O. Box 1450
Alexandria, Virginia 22313-1450
www .uspto.gov
ATTORNEY DOCKET NO. CONFIRMATION NO.
82923425 2282
EXAMINER
HICKS, MICHAEL J
ART UNIT PAPER NUMBER
2165
NOTIFICATION DATE DELIVERY MODE
05/06/2016 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):
hpe.ip.mail@hpe.com
mkraft@hpe.com
chris.mania@hpe.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte VINA Y DEOLALIKAR and HERNAN LAFFITTE
Appeal2014-007521
Application 13/383,592
Technology Center 2100
Before ALLEN R. MacDONALD, KEVIN C. TROCK, and
AMBER L. HAGY, Administrative Patent Judges.
MacDONALD, Administrative Patent Judge.
DECISION ON APPEAL
Appeal2014-007521
Application 13/383,592
STATEMENT OF THE CASE
Appellants appeal under 35 U.S.C. § 134(a) from a Final Rejection of
claims 1-16. We have jurisdiction under 35 U.S.C. § 6(b ).
Exemplary Claim
Exemplary claim 1 under appeal reads as follows (emphasis added):
1. A method of identifying a principal document in a document
set in a computer system, comprising:
obtaining a document set comprising a plurality of documents
from a storage device in the computer system;
grouping the plurality of documents into a plurality of clusters
based, at least in part, on a textual similarity between
each of the plurality of documents;
obtaining one or more descriptive terms corresponding to the
plurality of documents, wherein the descriptive terms are
terms within the plurality of documents that have been
identified as being useful for discriminating between
the clusters; and,
for each cluster:
identifying a subset of descriptive terms based, at least
in part, on a prevalence of the descriptive terms
within the documents of the cluster; and
identifying the principal documents in the cluster based,
at least in part, on a prevalence of the subset of
descriptive terms within each of the documents in
the cluster.
Rejections
The Examiner provisionally rejected claims 1--4, 6-8, 11-13, 15, and
16 on the ground of nonstatutory obviousness-type double patenting as not
2
Appeal2014-007521
Application 13/383,592
being patentably distinct from claims 1-20 of U.S. Patent Applications
12/705,584 and 12/705,586. Final Act. 4. 1
The Examiner rejected claims 1-16 under 35 U.S.C. § 103(a) as being
unpatentable over the combination of F. Archetti et al., A Hierarchical
Document Clustering Environment Based on the Induced Bisecting k-Means,
FQAS 2006, LNAI 4027, 257----69 (2006) and Anton Leuski, Evaluating
Document Clustering/or Interactive Information Retrieval, CIKM '01, 33-
40 (2001 ). Final Act. 4--11. 2
Appellants ' Contentions
1. Appellants contend that the Examiner erred in rejecting claim 1
under 35 U.S.C. § 103(a) because "Archetti does not disclose obtaining
descriptive terms that are useful for differentiating between clusters" (App.
Br. 8), particularly:
The Examiner argues that Sections 2.4, 2.5, and 2.6 of
Archetti discuss identifying the highest quality terms for a cluster
both for processes of clustering and sub-clustering and for
determining cluster labels. See Final Office Action, p. 5. The
Examiner also asserts that the terms are based on the quality of
the terms and include prevalence based metrics. See id. However,
the cited sections do not do so. Section 2.4 is related to preparing
the collection of documents for clustering. See Archetti, p. 260,
11. 21-23. Section 2 .4 only discusses calculating the weight of
each term based on Term Frequency (TF) and Inverse Document
1 Appellants do not appeal this rejection ("SOLE GROUND OF
REJECTION TO BE REVIEWED ON APPEAL" (App. Br. 5)).
Therefore, we affirm proforma. Except for our ultimate decision, this
rejection is not discussed further herein.
2 Separate patentability is not argued for claims 2-16. Except for our
ultimate decision, the Examiner's rejection of these claims is not discussed
further herein.
3
Appeal2014-007521
Application 13/383,592
Frequency (IDF), but not identifying which terms are useful for
cluster discrimination. See id., p. 260, 11. 27-31.
App. Br. 9 (emphasis omitted).
2. Appellants contend that the Examiner erred in rejecting claim 1
under 35 U.S.C. § 103(a) because "Archetti does not disclose ...
[identifying] a subset of the descriptive terms that are prevalent for each
cluster" (App. Br. 8), particularly:
[T]he Examiner argues that Term Frequency and Inverse
Document Frequency are used to select high quality terms. See
Final Office Action, pp. 2-3. However, Archetti does not disclose
identifying, for each cluster, a subset from the list of descriptive
terms. Rather, in regards to clusters, Archetti only describes the
manner in which clusters are split. See Archetti, p. 261.
Specifically, a cluster can be split based on two randomly
selected seeds. See id. The prevalence in each cluster of a subset
from a list of previously identified descriptive terms is not
considered in Archetti. Thus, Archetti does not disclose metrics
based on the prevalence of descriptive terms in each cluster.
App. Br. 9--10 (emphasis omitted).
3. Appellants contend that the Examiner erred in rejecting claim 1
under 35 U.S.C. § 103(a) because "Archetti does not disclose identifying
principal documents based on a prevalence of the subset of descriptive terms
for each cluster" (App. Br. 8-9), particularly:
The metrics described in Sections 2.4 and 2.5 [of Archetti] are
discussed only in relation to sorting documents into clusters,
rather than identifying principal documents from the clusters.
App. Br. 9.
4
Appeal2014-007521
Application 13/383,592
4. Appellants contend that the Examiner erred in rejecting claim 1
under 35 U.S.C. § 103(a):
Leuski does not disclose obtaining descriptive terms that are
useful for differentiating between clusters, as recited by the
independent claims. Nor does Leuski disclose identifying a
subset of the descriptive terms that are prevalent for each cluster,
or identifying a principal document for each cluster based on the
prevalence of the descriptive terms in the cluster.
App. Br. 10.
Issue on Appeal
Did the Examiner err in rejecting claim 1 as being obvious?
ANALYSIS
We have reviewed the Examiner's rejections in light of Appellants'
arguments (Appeal Brief and Reply Brief) that the Examiner has erred. We
disagree with Appellants' conclusions. Except as noted below, we adopt as
our own ( l) the findings and reasons set forth by the Examiner in the Action
from which this appeal is taken and (2) the reasons set forth by the Examiner
in the Examiner's Answer in response to Appellants' Appeal Brief. We
concur with the conclusions reached by the Examiner. We highlight the
following additional points.
As to Appellants' above contention 1, we disagree. We agree with the
Examiner that Archetti:
(1) searches and retrieves documents (Archetti Section 2.4);
(2) assigns weighted terms to each document where the terms used are
limited to terms with the highest quality (Archetti Section 2.4);
(3) performs hierarchal document clustering (Archetti Section 2.5);
( 4) computes a cluster centroid (Archetti Section 2.5); and
5
Appeal2014-007521
Application 13/383,592
(5) assigns centroid labels based on the top H weighted values of the
centroid vector by determining the terms which contribute to the top
H terms (Archetti Section 2.6).
We conclude that an artisan would understand this process to result in an
identification of the terms that were useful for cluster determination. Also,
contrary to any implication, in Appellants' argument, that the claimed
descriptive terms are used in the cluster creation step, such is not claimed or
disclosed. Rather, as with Archetti's centroid labels, the descriptive terms
merely serve to discriminate (distinguish) between existing clusters and are
determined after the clusters are generated. See Appellants' Figure 2.
As to Appellants' above contention 2, we disagree with Appellants'
argument for the reasons above and set forth by the Examiner (Ans. 11-12).
Separately, we note that an artisan would understand the contribution of
each term in Archetti is of different weight (e.g., "the terms with highest
quality," Archetti 261 ), and a subset of the terms in a cluster are the most
valuable to the process (e.g., "top H weighted values of the centroid vector,"
Archetti 261 ). Therefore, we conclude that identifying a subset of the terms
based on prevalence would have been obvious to an artisan.
As to Appellants' above contentions 3 and 4, we disagree with
Appellants' arguments for the reasons above and set forth by the Examiner
(Ans. 12-13). Further, these arguments ignore the actual reasoning of the
Examiner's rejection. Instead, Appellants attacks each reference singly for
lacking a teaching ("identifying principal documents based on a prevalence
of the subset of descriptive terms for each cluster," App. Br. 9) that the
Examiner relied on the combination of Archetti and Leuski to show.
6
Appeal2014-007521
Application 13/383,592
(See Section 3.3 [of Leuski] which clearly discloses that
representative documents for each cluster are chose, at that the
representative documents may be chosen based on the document
which is most similar to the centroid for the cluster. Examiner
notes that the centroids of Archetti are chosen based on the
prevalence of terms of the within the documents.).
Final Act. 6. It is well established that one cannot show nonobviousness by
attacking references individually where the rejections are based on
combinations of references. See In re Keller, 642 F.2d 413, 425 (CCPA
1981); In re Merck & Co., Inc., 800 F.2d 1091, 1097 (Fed. Cir. 1986).
Appellants are raising and then knocking down a straw man rejection that
was never made by the Examiner. This form of argument is unavailing to
show Examiner error.
CONCLUSIONS
(1) The Examiner has not erred in rejecting claims 1-16 as being
unpatentable under 35 U.S.C. § 103(a).
(2) The Examiner did not err in provisionally rejecting claims 1--4, 6-
8, 11-13, 15, and 16 on the ground of nonstatutory obviousness-type double
patenting as not being patentably distinct.
(3) Claims 1-16 are not patentable.
DECISION
The Examiner's rejection of claims 1-16 as being unpatentable under
35 U.S.C. § 103(a) is affirmed.
The Examiner's provisional rejection of claims 1--4, 6-8, 11-13, 15,
and 16 on the ground of nonstatutory obviousness-type double patenting is
affirmed.
7
Appeal2014-007521
Application 13/383,592
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 3
3 As the Examiner has shown that all the claims are unpatentable, we do not
also reject Appellant's claims 1-16 under 35 U.S.C. § 101 as not being
directed to patent eligible subject matter. Ho\x1ever, should there be fi1rther
prosecution of these claims; the Examiner's attention is directed to our
following concerns.
With respect to claims 1-16, the Examiner's attention is directed to
the Court's decision in Alice Corp. Pty. Ltd., v. CLS Banklnt'l, 134 S. Ct.
234 7 (2014) and any subsequent agency guidance. Subsequent to the close
of prosecution before the Examiner, the Supreme Court in Alice set forth a
two-part test to determine compliance of a claim with § 101. The Court held
that a mere instruction to implement an abstract idea on a computer cannot
impart patent eligibility. Alice Corp., 134 S. Ct. at 2357-59.
For example, claiming a process for clustering documents (themselves
consisting of abstract information content) according to an abstract
mathematical algorithm (no matter how allegedly novel the algorithm) is
unlikely to comply with 35 U.S.C. § 101 absent any "non-abstract" inventive
application of the formula. Note that application of the algorithm such as
steps of "identifying a subset of descriptive terms" or "identifying the
principal documents in the cluster" may not be sufficient to render a claim
statutory.
8