Ex Parte Golan et al

Patent Trial and Appeal Board

Jun 8, 2018

14354115 (P.T.A.B. Jun. 8, 2018)

UNITED STA TES p A TENT AND TRADEMARK OFFICE
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR
14/354,115 04/24/2014 Shahar Golan
56436 7590 06/12/2018
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.
83933546 5939
EXAMINER
EDWARDS, JASON T
ART UNIT PAPER NUMBER
2144
NOTIFICATION DATE DELIVERY MODE
06/12/2018 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
chris.mania@hpe.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte SHARAR GOLAN, OMER BARKOL,
RUTH BERGMAN, IRA COHEN, and
GAL NOY
Appeal2017-010882
Application 14/354,115
Technology Center 2100
Before CAROLYN D. THOMAS, CARL W. WHITEHEAD JR. and
ADAM J. PYONIN, Administrative Patent Judges.
WHITEHEAD JR., Administrative Patent Judge.
DECISION ON APPEAL
STATEMENT OF THE CASE
Appellants are appealing the final rejection of claims 1-10 and 12-21
under 35 U.S.C. § 134(a). Appeal Brief 1. We have jurisdiction under 35
U.S.C. § 6(b) (2012).
We reverse.
Introduction
The invention is directed to:
A method and a system for aligning the annotation of fields of
documents are provided, according to various embodiments. For
example, a plurality of documents that belong to the same type
of document are used to generate an electronic classification
Appeal2017-010882
Application 14/354,115
model based on training information obtained by analyzing the
documents. Examples of document types include news, games,
invoices and information technology. In this case, documents for
CNN, BBC and New York times are examples of documents that
are of the same type since they are all news type documents.
The fields associated with a training set of documents are
assigned to clusters, according to various embodiments, so that
each cluster is for a type of field. For example, there may be a
first training cluster for user/author/poster type fields, second
training cluster for title/subject type fields, and third training
cluster for message/body/post type fields. The clusters can then
be annotated according to their field type by a human. For
example, the human may annotate the first training cluster with
"author," the second training cluster with "title," and the third
training cluster with "message."
The features for each of the fields associated with the training
clusters can be measured. The measurements of the features of
the fields for each of the clusters are also referred to as "training
information." A learning based classifier, according to one
embodiment, receives the training information and generates an
electronic classification model based on the training information.
Specification i-f i-f 11, 15, 16, respectively.
Illustrative Claim
1. A method comprising:
accessing training information that includes first measurement
information pertaining to features of a plurality of fields associated with
training clusters for documents of a document type, the training clusters
produced from clustering the plurality of fields in the documents into
respective different training clusters for respective different field types,
wherein a first training cluster associated with a first field type of the
different field types is annotated with a first name and comprises fields of
the first field type from the documents, and a second training cluster
2
Appeal2017-010882
Application 14/354,115
associated with a second field type of the different field types is annotated
with a second name and comprises fields of the second field type from the
documents;
generating an electronic classification model based on the training
information;
accessing second measurement information for features of fields
associated with new clusters of a new document; and
automatically annotating, performed by a computer system, each of
the new clusters based on the second measurement information using the
classification model, wherein a first new cluster that has fields of the first
field type is annotated with the first name, and a second new cluster that has
fields of the second field type is annotated with the second name.
Rejections on Appeal1
Claims 1, 5, 6, 7, 10, 12, 15, 16, 18 and 21 stand rejected under pre-
AIA 35 U.S.C. § 103(a) as being over Nevill-Manning (US Patent 7,505,984
B 1; issued March 17, 2009) and Knight (US Patent Application Publication
2011/0029525 Al; published February 3, 2011). Final Action 3-12.
Claims 2 and 20 stand rejected under pre-AIA 35 U.S.C. § 103(a) as
being over Nevill-Manning, Knight and Ciemiak (US Patent 8,374,975 Bl;
issued February 12, 2013). Final Action 13-15.
Claim 3 stands rejected under pre-AIA 35 U.S.C. § 103(a) as being
over Nevill-Manning, Knight, Ciemiak and Zheng (US Patent Application
Publication 2011/0191381 Al; published August 4, 2011). Final Action 15-
16.
Claims 8, 9, 13, 14 and 17 stand rejected under pre-AIA 35 U.S.C. §
103(a) as being over Nevill-Manning, Knight and Zheng. Final Action 16-
19.
1 Appellants state the Examiner's"§ 112, i-f 2, rejection of claim 6 has been
rendered moot" due to the claim amendment submitted January 20, 2017.
Appeal Brief 6.
3
Appeal2017-010882
Application 14/354,115
Claim 4 stands rejected under pre-AIA 35 U.S.C. § 103(a) as being
over Nevill-Manning, Knight, Ciemiak and Younes (US Patent 8,589,366
Bl; published November 19, 2013). Final Action 19-20.
Claim 19 stands rejected under pre-AIA 35 U.S.C. § 103(a) as being
over Nevill-Manning, Knight and Younes. Final Action 20-21
ANALYSIS
Rather than reiterate the arguments of Appellants and the Examiner,
we refer to the Appeal Brief (filed March 20, 2017), the Reply Brief (filed
August 22, 2017), the Answer (mailed June 23, 2017) and the Final Action
(mailed October 20, 2016) for the respective details.
Appellants contend the following:
the obviousness rejection of claim 1 is erroneous for at least two
reasons: (1) the Examiner erred in asserting that Nevill-Manning
teaches the subject matter in the "accessing training information"
clause of claim 1, and (2) the Examiner erred in asserting that
Knight teaches subject matter in the "accessing second
measurement" and "automatically annotating" clauses of claim
1.
Appeal Brief 7.
We note Nevill-Manning discloses aligning the annotation of fields of
documents similar to the claimed invention:
Methods and systems for information extraction are disclosed.
In one such method and system, a sample of related articles is
obtained, and an article is selected as a seed article. The distances
between sample articles are calculated to determine a set of one
or more closest articles to the seed article. The set of closest
articles is used to identify information fields containing variable
data within the seed article. There are a variety of techniques by
which this may be performed, one of which is by using dynamic
programming alignment to compute alignments between articles.
The information fields are labeled, and a template is generated
4
Appeal2017-010882
Application 14/354,115
using the labeled fields. The template is used to extract data from
a source article by comparing the source article with the template
and associating the variable data of the source article with the
labeled fields.
Nevill-Manning, Abstract.
The Examiner finds Nevill-Manning discloses, "[f]orming clusters of
fields across documents within a collection of documents (col. 7, lines 24--
29; Abstract; Fig. 3; col 7, lines 1-7; col. 5, lines 37--40; col. 8, lines 28-
34)." Answer 22. Appellants argue:
Fundamentally, what occurs in Nevill-Manning is the
identification [of] certain fields for inclusion in a template, and
assigning labels to such individual fields in the template. This is
quite different from producing training clusters for different field
types where each training cluster includes multiple fields and
each training cluster is assigned a respective name, as claimed.
Appeal Brief 9.
Appellants contend:
Among the various passages cited by the Examiner's Answer,
Nevill-Manning refers to determining constant data portions of a
seed document and variable data portions of the seed document.
Nevill-Manning, 7: 17-19. Where the data on the documents
differs from the data on the seed document indicates variable data
and where the data on the documents is the same as the data on
the seed document indicates constant data. Id., 7:21-24.
Reply Brief 4--5.
Appellants argue, "Nevill-Manning refers to identifying 'variable
portions of the documents' 'as information fields.' Nevill-Manning, 7:39-41.
The information fields are labelled, such as by a 'template generator [that]
automatically generates labels for relevant information fields.' Id., 7 :41, 55-
56." Appeal Brief 8. Appellants conclude, "Thus, Nevill-Manning refers to
comparing documents to a seed document to determine which data portions
are constant and which data portions are variable. No reference is made to
5
Appeal2017-010882
Application 14/354,115
producing training clusters of fields of respective different types." Reply
Brief 5.
We find Appellants' arguments persuasive and disagree with the
Examiner's findings. As we have noted above, both the claimed invention
and Nevill-Manning aligns the annotation fields of documents, however
Nevill-Manning fails to disclose training clusters of fields for documents as
claimed. Knight fails to address Nevill-Manning's deficiency because
Knight teaches "that the training set of known results [0023] can be used to
provide suggestions for classifications of a set of uncoded documents in a
corpus [0032]." Final Action 5.
Accordingly, we reverse the Examiner's obviousness rejection of
independent claims 1, 6 and 12, as well as, dependent claims 5, 7, 10, 15, 16,
18 and 21. We also reverse the Examiner's obviousness rejections of
dependent claims 2--4, 8, 9, 13, 14, 17, 19 and 20 because Ciemiak, Zheng
and Younes fail to address the noted deficiency of Nevill-Manning.
DECISION
The Examiner's obviousness rejections of claims 1-10 and 12-21 are
reversed.
REVERSED
6