Ex Parte 6778979 et al

Patent Trial and Appeal Board

Jan 31, 2014

95001430 (P.T.A.B. Jan. 31, 2014)

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.
95/001,430 09/08/2010 6778979 19473-0036RX1 1480
11580 7590 02/03/2014
Schwegman Lundberg & Woessner/Xerox
P.O. Box 2938
Minneapolis, MN 55402
EXAMINER
WOOD, WILLIAM H
ART UNIT PAPER NUMBER
3992
MAIL DATE DELIVERY MODE
02/03/2014 PAPER
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.
PTOL-90A (Rev. 04/07)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
____________

GOOGLE, INC.
Requester and Respondent

v.

XEROX CORPORATION
Patent Owner and Appellant
____________

Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2
Technology Center 3900
____________

Before RICHARD M. LEBOVITZ, DAVID M. KOHUT, and
ANDREW J. DILLON, Administrative Patent Judges.

LEBOVITZ, Administrative Patent Judge.

DECISION ON APPEAL
This is a decision on the appeal by the Patent Owner from the Patent
Examiner’s decision to reject pending claims 1-20 in the above-identified inter
partes reexamination of U.S. Patent No. 6,778,979 B2. The Board’s jurisdiction
for this appeal is under 35 U.S.C. §§ 6(b), 134, and 315. We affirm and set forth a
new ground of rejection.
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

2

BACKGROUND
The patent in dispute in this appeal is US 6,778,979 B2 (“the ‘979 patent”)
which issued August 17, 2004. The claims are directed to methods for
automatically generating a query from a document using classification labels and
“entities” in the document content, where an entity is something recognized in a
document, such as a name, location, medical term, etc. (‘979 patent, col. 6, l. 65 to
col. 7, l. 2). A request for inter partes reexamination of the ‘979 patent under 35
U.S.C. §§ 311-318 and 37 C.F.R. §§ 1.902-1.997 was filed August 27, 2010 by
Google (Request 1). The real party in interest is Xerox Corporation (Appeal Br. 1,
filed January 2, 2013).
The ‘979 patent was previously the subject of litigation, in the U.S. District
Court for the District of Delaware, case no. 1:10-cv-00136-UNA (Appeal Br. 1).
Third Party Requester was a defendant in that litigation. According to Patent
Owner, the litigation was settled as to Third Party Requester as of November 10,
2011; and was settled as to the remaining party to the case as of December 12,
2011 (id.). Patent Owner stated that Third Party Requester withdrew from any
further participation in the reexamination, including in this appeal, as of November
30, 2011 (id.).
Fourteen rejections of the claims have been set forth by the Examiner as
discussed in more detail below. Although all 20 claims have been rejected under
the different rejections, Patent Owner only argued independent claims 1, 14, and
18 for six of the rejections, and argued in detail only with respect to claim 1
(Appeal Br. 4-5).
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

3

CLAIM 1
1

1. A method for automatically generating a query from selected
document content, comprising:
[1] defining an organized classification of document content
with each class in the organized classification of document content
having associated therewith a classification label;
each classification label corresponding to a category of
information in an information retrieval system;
[2] automatically identifying a set of entities in the selected
document content for searching additional information related thereto
using the information retrieval system;
[3] automatically categorizing the selected document content
using the organized classification of document content for assigning
the selected document content a classification label from the
organized classification of content; and
[4] automatically formulating the query to restrict a search at
the information retrieval system for information concerning the set of
entities to the category of information in the information retrieval
system identified by the assigned classification label.

The first step of the claim is “defining an organized classification of
document content.” The organized classification involves defining “classes” in the
document, where each class has a “classification label” that is defined in the claim
as “corresponding to a category of information in an information retrieval system.”
An information retrieval system can be Google or Yahoo (‘979 patent, col. 49, ll.
49-51).
According to the ‘979 patent specification, the classification is performed by
a “text categorizer 3610” based on content (words, sentences, etc.) which occur in
the document (id. at col. 43, l. 35 to col. 44, l. 56; col. 48, l. 52-55; col. 49, ll. 17-

1
Numbered steps [1] to [4] have been added for reference.

Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

4

48). The ‘979 patent does not define “classification label,” but the patent describes
the labels as being generated by a “categorizer” and makes clear that the labels
correspond to categories (id. at col. 49, ll. 18-30). Examples of classes or category
labels are class hierarchies, such as “science>biology>genetics” (id. at col. 49, ll.
22-26; col. 50, ll. 8-11). The patent also distinguishes between categories and
keywords (id. at col. 68, l. 23). Thus, a “classification label” appears to be used in
the patent in accordance with the ordinary meaning of “class” as a “set, collection,
group, or configuration containing members regarded as having certain attributes
or traits in common,”
2
and does not simply correspond to the text, keywords,
graphics, etc., that appear in the document.
One of the issues involved in the interpretation of claim 1 is whether the
“classification labels” have to be predefined or preexisting. Since subsequent
steps, such as steps [3] and [4] reference the “classification label” defined in step
[1], we interpret step [1] as establishing the “classification labels” to be used in
subsequent steps [3] and [4]. The “classification labels” are therefore predefined in
step [1]. The “classification labels” also correspond to “a category of information
in an information retrieval system” and thus represent “preexisting” categories in
an information retrieval system.
Step [2] involves “automatically identifying a set of entities in the selected
document content for searching additional information.” The term “entity” is
defined in the ‘979 patent to mean “something recognized in a document . . . that
can be in the form of an image, text, embedded data, HTML, etc.” (id. at col. 6, l.
65 to col. 7, l. 2).

2
http://www.thefreedictionary.com/class
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

5

Step [3] comprises “automatically categorizing the selected document
content using the organized classification of document content for assigning the
selected document content a classification label from the organized classification
of content.” In other words, the selected document is categorized by assigning it a
classification label from the organized classification defined in step [1] of the
method. Step [3] must therefore come after step [1] in the claim. The assignment
is based on the document content. As explained in the ‘979 patent:
F.1 Text Categorizer
The goal of a text classification system, such as text categorizer 3610,
is to classify a document 3612 into a set of one or more classes 3620,
which are also referred to as categories. In operation, the text
categorizer 3610 assigns a document one or more classes in a set of
classes that are defined in an ontology represented in knowledge base
3622. An example of an ontology is the DMOZ ontology (published
on the Internet at dmoz.org).
(Id. at col. 41, ll. 52-60)
The ‘979 patent describes using features of the document, such as words in
the document (id. at col. 43, ll. 61-63) to perform the “automatically categorizing
the selected document content.” Various algorithms for performing the
categorization are described in the patent (id. at col. 41, l. 52 to col. col. 46, l. 67;
probabilistic model, fuzzy model, vector space model).
Step [4] comprises “automatically formulating the query.” The query is
formulated to “restrict” the search relating to the “set of entities” to the category of
information identified by the classification label. Thus, step [4] must follow steps
[1], [2], and [3] because it uses the label of step [1] and [3] and the entities of step
[2].

Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

6

1. READER
Claims 1, 5, 10, 13, 14, and 18 stand rejected as anticipated under 35 U.S.C.
§ 102(e) by Reader
3
(RAN 5). Reader describes an automated search technique for
discovering patent-relevant publications on the Internet (Reader, Abstract). Patent-
identifying attributes, such as an inventor name, assignee’s name or patent number,
are used to retrieve a patent classification (PC) and patent language (PL) from a
patent database (page 3, col. 1, ll. 12-20; page 1, 4 ¶). The patent classification
attribute is used to form a company website identifying search query which used to
find company websites (id. at page 3, col. 1, ll. 17-37). The documents on the
website are subsequently searched (id. at page 3, col. 1, ll. 47-56).
Patent Owner contends that Reader does not perform steps [3] and [4] of the
claimed method.
According to the Requester, step [3] “automatically categorizing” the
document content is accomplished in Reader by submitting a search query using
patent attributes to a patent server which then retrieves a patent classification (PC)
and patent language (PL) as described in paragraph 15 (Request 37-38).
We agree with Patent Owner that Reader’s description of retrieving a patent
classification from a patent database is not “automatically categorizing” document
content as that step would be understood by one of ordinary skill in the light of the
‘979 patent. As discussed in the claim interpretation section, the automatic
categorizing is a step in which the content is analyzed and assigned a category
based on its content, e.g., using a specific algorithm, such as probabilistic, fuzzy,
and vector space models (‘979 patent, col. 41, l. 51 to col. col. 46, l. 67). In
Reader, the classification label is not obtained by categorizing the content of the

3
Reader, US 2002/0147738 A1, published October 10, 2002.
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

7

document, but rather is obtained from a database by using the patent attribute to
retrieve an already existing classification label associated with the patent attribute.
Retrieving the label from a patent database is not the same operation as
automatically categorizing because it is not using the content of the document to
assign a label. Rather, it is using the patent attribute of the document, such as an
inventor name, assignee name or patent number, as a tag to retrieve an already
existing classification associated with the tag in the database.
The Examiner in adopting the Requester’s proposed rejection takes the
position that the categorizing is assigning a classification to the document (RAN
31), but does not explain how retrieving a label from an already classified
document meets the limitation of “categorizing” as that term would be understood
in light of the ‘979 patent. Categorizing is not recited alone in the claim, but rather
the categorizing is with respect to the “document content.” If Reader is understood
to perform categorizing by retrieving a class from a database, the categorizing is
not accomplished using the document content, but rather is performed using a
patent attribute which is defined by Reader to include an inventor name, assignee
name or patent number (Reader 4). These “attributes” are not representative of the
document content, but are more like tags assigned as identifiers to document, not
the content of the document.
We also agree with the Patent Owner that Reader does not teach step [4] of
the claimed method. Step [4] requires the search to be restricted to a category of
information identified by the classification label. Requester argued that the
classification label is the patent classification retrieved from the patent database
(Request 37). For this step, Requester refers to Reader’s discussion at paragraph
15 which describes (1) using the patent classification (PC) to form a company
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

8

website search (CWI) on patent; and (2) “[u]sing the company website identifier
and well known DNS addressing, . . searches the totality of full-text documents
published on the associated company website for Web document language relevant
to the WDI attributes” (Reader, ¶ 15; quoted on page 39 of Request). Thus,
although the search is restricted as required by step [4] of the claimed method, the
restriction is not to the category of information identified by the patent
classification label. Rather, the patent classification label is used (1) to search the
entire database for a company without any restriction. Subsequently, (2) the search
is restricted to a company website, but this search is restricted by the company
identifier, not the patent classification label as required by the claim.
The Examiner states that “Reader is using the patent classification to
determine the companies involved and retrieve their web addresses so that their
web documents may be searched. Reader uses the classification/categorization to
restrict to searching certain company web documents.” (RAN 36.) The disclosure
of Reader does not support the Examiner’s statements. As explained above,
Reader uses “the company website identifier [not the patent classification] and well
known DNS addressing” to search “the totality of full-text documents” (Reader, ¶
15; quoted on page 39 of Request). Thus, the preponderance of the evidence does
not support the Examiner’s finding that step [4] of the claimed is met by Reader.
For the foregoing reasons and those of the Requester’s, we reverse the
rejection of claim 1. The rejection of claims 5, 10, 13, 14, and 18 are reversed for
the same reasons.
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

9

2. READER AND MASE
Claims 1, 5, 10, 13, 14, and 18 stand rejected as obvious under 35 U.S.C.
§ 103(a) in view of Reader and Mase
4
(RAN 5). Reader describes an automated
search technique for discovering patent-relevant publications on the Internet
(Reader, Abstract). Mase is cited for step [3] of automatically categorizing
documents (Request 50). With regard to step [4], Requester cites the same
disclosure in Reader as they did for the rejection based on Reader as an
anticipatory publication (id. at 53-54). Requester contends that person of ordinary
skill in the art would have recognized that “Reader's techniques could be
advantageously combined with Mase's techniques to achieve a system that would
provide an examiner at the patent office with both a patent application that is
automatically classified in accordance with Mase's teachings and Web documents
automatically determined as potential prior art relevant to the patent application in
accordance with Reader's teachings.” (Request 53)
Requester provided a logical reason for combining the publications.
However, the rejection is defective because – as explained above in the section on
Reader alone – step [4] is not described by Reader. The Examiner’s position on
step [4] involves the same unpersuasive arguments as for Reader, alone (RAN 41-
44).
The rejection of claims 5, 10, 13, 14, and 18 are reversed for the same
reasons.

4
Mase, “Experimental Simulation for Automatic Patent Categorization” IFIP
1996.
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

10

3. WIESER
Claims 1, 5, 6, 10-14, and 18 stand rejected as anticipated under 35 U.S.C.
§ 102(a) by Wieser
5
(RAN 6). Wieser describes a method and system for
automatically contextually matching offers to documents on a web page that is
selected by a user in real-time (Wieser, Abstract). As explained in the publication:
For example, if a visitor or customer of client 12 selects an article
about BMW's newest vehicle, the match server 14 of the present
invention presents a buying opportunity to the customer, i.e., an
attractively discounted book for “BMW Enthusiasts”. If the customer
is reading an article about Tiger Woods, the match server 14 displays
or presents items such as Nike golf balls or a book on the history of
the US Open to the customer.
(Wieser 6: 4-9.)
Patent Owner disputes the Examiner’s findings that Wieser describes steps
[1], [3], and [4] of the claimed method (Appeal Br. 24-25).
Step [1]. Requester states Wieser teaches “clusters” which serve as classes.
Requester’s identification of “clusters” is confused. First, Requester cites to
Wieser’s disclosure of grouping related “documents” (Requester 61). Second,
Requester refers to Wieser’s discussion of clusters of “products (id. at 61-62).
These are two different, unrelated teachings.
Step [3]. Requester refers to the product clusters as classification labels, and
states that Wieser performs “automatically categorizing” documents by “assigning
the selected document one or more cluster IDs from the organized classification of
document content.” (Request 63.) Requester explains that the selected document
is automatically categorized “by the match server 14 into one or more clusters of
content having corresponding cluster IDs by matching the query context vector of

5
Weiser, WO 01/44992 A1, published June 21, 2001.
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

11

the selected document with the one or more spheres or clusters in vector space.”
(id.) The Wieser clusters referenced by Requester are clusters of products:
After the vector generations, the document vector, or feature
vector, must be compared with database vectors, or feature vectors. A
naïve approach to product matching would be to compare the
document vector to the vectors for every product in the database. This
process becomes more burdensome as the number of products in the
database grows. Fortunately, the database vectors are not smoothly
distributed throughout the vector space, but rather, tend to "clump"
together, leaving vast empty spaces between the clumps, or clusters as
they are commonly known.
. . . Then, when one wishes to find the products that match a given
document vector, one need only compare the document to the
products in the cluster whose bounding sphere contains the document
vector (or the nearest spheres, if no sphere contains the document
vector).
(Wieser, p. 17, 15-28; emphasis added.)
Step [4]. In the section discussing this step, Requester states that “the
matching process performed by the match server 14 automatically forms an SQL
query that restricts the search of the inventory of product offers to those product
offers having product vectors within only the clusters having cluster IDs assigned
to the selected document content.” (Request 63-64.) In other words, once the
document is assigned a product cluster, a query is formulated to only search the
products in the cluster (id. at 63).

Discussion
We agree with Patent Owner that Wieser’s disclosure of keywords does not
describe the first step of the claimed method. The first step pertains to “defining
an organized classification of document content with each class in the organized
classification of document content having associated therewith a classification
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

12

label.” For the organized classification, Requester identifies the disclosure at page
15, lines 6-10, of Wieser:
According to one embodiment, the present invention provides a novel
approach to representing textual documents as high dimensional
vectors. Such an approach provides an efficient means of indexing
document collections, allowing retrieval of documents (querying)
based on keywords, grouping related documents (categorization).
Any grouping is based on keywords. Wieser teaches that its vector based
algorithms “all characterize documents based on the presence of keywords.”
(Wieser, p. 15, ll. 2-4.) Requester did not explain how “keywords” are labels that
correspond to categories of information as required by the claim. The ‘979 patent
distinguishes between categories and keywords (‘979 patent, col. 68, l. 23).
Alternatively, Requester characterized the product clusters as labels. This
explanation is more consistent with how Requester found the elements of steps [3]
and [4] in the claim (see above). In this explanation, entities would be picked out
of the document (step [2]) and used to search the product in step [4] after narrowed
by the product clusters in step [3].
Patent Owner contends that the product clusters not labels or categories
because they are grouped by mathematical formula which may not belong to a
“coherent human-understandable category,” quoting from the declaration by Dr.
Subrahmanian (Appeal Br. 30; Declaration of Dr. V.S. Subrahmanian,
6
dated June
7, 2011). However, even if this is true, the claim does not require the categories to
be human understandable, as long as they can be used in subsequent categorizing
and formulating steps. Patent Owner describes how the documents are classified in

6
Dr. Subrahmanian states that he has a Ph.D. in computer science and has written
papers on information management, ontologies, and textual informational
processing (Subrahmanian Decl. 2, 6, and 9) and thus has experience relevant to
the field of the claimed invention in this proceeding.
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

13

the ‘979 patent (id. at 30-32), but the claim is not limited to this method and it is
improper to import limitations from the Specification into the claims.
With respect to step [3], Patent Owner asserts Wieser is deficient because
the product clusters are not classification labels (Appeal Br. 33). However, as
explained above, we do not find this argument persuasive.
Dr. Subrahmanian testified:
I do not find any disclosure in Wieser that would identify how the
clusters in his patent application are mapped to such hierarchical
representations. It is not apparent from Wieser how such mapping
from Wieser’s described clusters to hierarchical representations might
be performed.
(Supplemental Declaration of Professor V.S. Subrahmanian ¶ 21.)
The claim does not require “hierarchical representations.” It is unclear what
step of the claimed method that Dr. Subrahmanian finds to be deficient in Wieser.
As explained above, Requester explained cogently how steps [3] and [4] are met.
Requester’s explanation is supported by a preponderance of the evidence. Patent
Owner does not clearly identify any defect in Requester’s facts, and we find none
either, aside from Requester’s faulty but unnecessary reliance on document
grouping in step [1].
On page 17 of Wieser, Wieser describes matching document vectors to the
product clusters, where this disclosure corresponds to step [3]. Page 18 of Wieser
explains how this is accomplished. It not clear from Dr. Subrahmanian what is
defective about this disclosure, and why it fails to describe and enable step [3] of
claim 1.
For the foregoing reasons and those of the Requester as adopted by the
Examiner, we affirm the rejection of claim 1 as anticipated by Wieser. Claims 5,
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

14

6, 10-14, and 18 have substantially the same limitations as claim 1, and were not
argued separately, and therefore fall with claim 1. 37 C.F.R. § 41.37(c)(1)(VII).

4. RHODES
Claims 1, 2, 5-7, 10-15, 18, and 19 stand rejected as anticipated under 35
U.S.C. § 102(a)/102(e) by Rhodes
7
(RAN 7). Rhodes describes a system in which
documents are stored in a database and searched for relevance to contextual
information instead of only similar text (Rhodes, Abstract). Patent Owner
contends that Rhodes does not describe all the steps of the claimed method (Appeal
Br. 35-38).
According to Requester, the classification labels of the document content [1]
are the “wordcodes” set forth in a “wordvec” document (Request 79). Requester
states that words are extracted by Rhodes from a document for additional searching
(id. at 80) as step [2] of the claim. Requester contends that [3] “automatically
categorizing the selected document content” is accomplished by identifying words
appearing in the selected document and then determining their appearance in a
“wordvec” file. Requester explains:
For example, a selected document including the word "pickle" is
matched against and produces the numeric wordcode "12". The
wordcode "12" then yields a vector, revealing all documents that
include at least one occurrence of the corresponding word "pickle"
against which more complex processing can be performed. In this
process, derivation of the wordcode "12" for the word "pickle," is
done for the selected document, such that the selected document is
automatically categorized.
(Request 81.) Requester cites column 5, ll. 12-28 for this disclosure (id. at 82).

7
Rhodes, US 6,236,768 B1, issued May 22, 2001.
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

15

With respect to the step [4] of “automatically formulating the query,”
Requester relies on Rhodes description of “automatically identifying one or more
wordcodes for the selected document and performing a similarity analysis that
restricts the documents in the corpus that are analyzed to only those documents that
are within the word vector group or groups corresponding to the wordcode(s),”
citing to the same passage at column 5, lines 12-28, cited for step [3].
Rhodes teaches creating a document vector of all the words and their
frequency which appear in a selected document and also against other documents
in a corpus of documents (Rhodes, col. 2, ll. 25-53; col. 3, ll. 1-31). Rhodes
describes creating a “wordvec file” which “contains each word appearing in the
entire indexed corpus of documents followed by a list of each document that
contains that particular word” (id. at col. 4, ll. 22-27). The wordvec file has
numeric wordcodes (WORDCODE-1, WORDCODE-2, etc.) which correspond to
the words in a document (id. at col. 4, ll. 45-46: “Each word in the wordvec is
represented by a unique numerical code”). Word queries are generated from a
sample text (id. at col. 1, l. 66 to col. 2, l. 5) and then matched to the wordvec file,
which generates a list of matching files based on the word similarities (id. at col. 5,
ll. 12-17).
Requester has relied upon the same disclosure in Rhodes for steps [3] and
[4]. In other words, Requester has stated that [3] categorizing of the document is
done when a word query is made of the wordvec file, but Requester relied on this
same disclosure for step [4] of automatically formulating query. Requester has not
given a reason for interpreting the steps to be the same. Step [3] involves
“automatically categorizing” by assigning a classification label to the document.
Step [4] involves “automatically formulating the query to restrict the search” to
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

16

“the category of information . . . identified by the assigned label.” Step [4] thus
requires step [3] of assigning the label before carrying out step [4]. Thus, the steps
are distinct and both are required to meet all the limitations of the claim.
Requester has not clearly identified disclosure in Rhodes where each of these steps
is independently described. The Examiner simply argues the breadth of the claim,
but does not persuasively point to disclosure in Rhodes which describes steps [3]
and [4] as independent steps (RAN 57-60). Consequently, a preponderance of the
evidence does not support the findings that Rhodes describes all the elements of
claim 1. The rejection can therefore not be sustained.
The rejection of claim 1 as anticipated by Rhodes is reversed. Claim 2, 5-7,
10-15, 18, and 19 have substantially the same limitations, and are reversed for the
same reasons.
5. BLACK AND HYPURSUIT
Claims 1, 2, 4-7, and 10-20 stand rejected as obvious under 35 U.S.C.
§103(a) in view of Black
8
and HyPursuit
9
(RAN 9). Patent Owner contends that
the combination of Black and HyPursuit do not teach or suggest steps [3] and [4]
of the claimed methods (Appeal Br. 39-44).
Requester cited HyPursuit for teaching the step [1] of defining an organized
classification of document content involving classification labels corresponding to
a category of information (Request 101-102). HyPursuit teaches clustering
contents and assigning content labels that summarize the cluster, where the clusters
can be based on different principles of grouping of documents, such as the Library

8
Black, US 6,546,386 B1, issued April 8, 2003.
9
Weiss, “HyPursuit: A Hierarchical Network Search Engine that Exploits Content-
Link Hypertext Clustering” ACM (1996) (hereinafter “HyPursuit”).
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

17

of Congress catalog scheme (Request 101; HyPursuit, p. 180). Patent Owner did
not challenge this argument, and we find it supported by the evidence.
With respect to step [2] of “automatically identifying a set of entities in the
selected document content,” Requester cites to Black’s disclosure of identifying
keywords in a body of content (Request 102; Black, col. 2, ll. 41-56). Black
teaches that “[k]eywords are simply a collection of words, generated automatically
. . . , that are deemed to be indicative of the topic matter or one of the topics for a
given content selection.” (Black, col. 2, ll. 26-29.)
Step [3] of automatically categorizing the selected document content by
assigning classification label is said by Examiner and Requester to be met by
Black’s description of a hook (Request 104). Black teaches:
The hook is the concept, primary subject matter or main topic for a
body of text. The hook is used to define a query as narrowly as
possible on a particular topic for a selected information source. To
determine a “hook”, a content layer must exist for which a context can
be determined. There must be a perceivable structure to the
information source and each content entry must have an associated
context or place or places within the structure of the information
source.
(Black, col. 2, ll. 10-18.)
In one of the examples, Black describes automatically generating a hook by
extracting proper names from a document and then, based on frequencies and other
context, identifying a proper name in the document to use as hook (Black, col. 2, l.
57 to col. 3, l. 24). The classification label assigned to the document in this case is
the proper name.
For step [4] of “automatically formulating the query,” Black teaches:
Once the keywords have been selected [step 2 of the claim] and the
hook for a body of text has been determined or automatically
generated [step 3 of the claim], the searches are created by generating
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

18

a link for every keyword extracted from the body of text and
combining it with the hook in a search that results in a result set that is
the logical intersection of the results generated by the hook and the
keyword.
(Black, col. 3, ll. 37-43.) The search can be performed on the Internet using a
search engine (id. at col. 2, ll. 23-25).
While Black describes [4] formulating a query as in the claimed method,
Black does not describe restricting the search to a category of information that is
identified by the classification label (the hook) assigned in step [3]. However, the
Examiner and Requester cited HyPursuit for this feature. HyPursuit describes
automatically categorizing a document based on its content:
The abstraction function for query routing, on the other hand,
computes a manageable set of terms that are used for identifying
portions of the information space relevant to particular queries. The
abstraction function uses term and term frequency information in the
children's content labels to compute term weights. The abstraction
function then selects the most heavily weighted terms for generating
the content router's content label.
(HyPursuit, p. 185, col. 1; emphasis in original.)
HyPursuit also describes restricting the search based on the assigned
labels:
HyPursuit uses query routing to support the search operations. Query
routing uses the content labels stored in the content router to
determine which of the child servers are likely to contain documents
related to the user query.
(HyPursuit, p. 186, col. 2.)

Arguments
Patent Owner contends that Black does not perform step [3] of the claimed
method:
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

19

Black's generation of the hook is done without any knowledge of an
organized classification of document content allegedly defined in the
combined system by the contribution from HyPursuit. Neither the
Request[er] nor the Examiner provide any explanation regarding how
the combined system of Black and HyPursuit would actually use the
organized classification of document content to assign the selected
document content a classification label when the only method of
generating the hook cannot guarantee that the hook is a classification
label within the organized classification of document content defined
by HyPursuit. Black's hook is nothing more than a proper noun from
the document, it is present only by being a noun in the document, and
is therefore not a classification label from any organized classification
of content.
(Appeal Br. 40 (emphasis in original); see also Appeal Br. 44 discussing lack of
motivation to combine Black and HyPursuit.)
This argument is not persuasive. Black does not limit its hook to proper
names, but rather defines hook broadly as “the concept, primary subject matter or
main topic for a body of text.” (Black, col. 2, ll. 10-11.) The generation of a hook
as “proper name” is exemplified by Black, but the evidence does not show that the
hook is restricted to proper names. Black teaches the concept of generating a hook
from the content of a body of text to “to define a query as narrowly as possible on
a particular topic for a selected information source.” (Id. at col. 2, ll. 10-18.) Such
disclosure reasonably suggests using HyPursuit’s method of clustering content for
the purpose of improving searching (HyPursuit, “Document clustering” and
“Hierarchical Organization of a HyPursuit System, p. 183) because HyPursuit’s
clusters can correspond to topics such as “video databases,” Library of Congress
catalog subjects, literature, and “predefined categories, including technology,
health, business, weather, and others” (id. at p. 183, col. 2; p. 184, col. 1; p. 185,
col. 2; and p. 190, col. 2), just as Black’s hooks represent topics.
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

20

Patent Owner contends that “the method used by Black to obtain the ‘hook’
is not described as using a predefined categorization system, and could not without
non-obvious modification use a pre-defined categorization system.” (Appeal Br.
41).
This argument does not persuade us that the Examiner erred. Black
generally suggests uses a categorization system based on topics. Based on this
general teaching, one of ordinary skill in the art would have reasonably found
HyPursuit’s teaching of using predefined categories in Black’s method obvious as
matter of using HyPursuit to improve searching capability. Black does not restrict
its hook, but rather defines it broadly to encompass the predefined categories in
HyPursuit, since the latter are “topics” and thus meet Black’s definition of a hook.
Patent Owner also contends that combination of Black and HyPursuit fails to
teach or suggest the automatically formulating the query element of the claimed
method (Appeal Br. 42). According to Patent Owner, the query is not automatic in
HyPursuit, but operates only in response to user selection (id.)
This argument is not persuasive. Black is relied upon for automatically
formulating a search based on the keywords and hooks. Black describes using
keywords and hooks to automatically search a database (Black, col. 3, ll. 37-43).
Thus, even if HyPursuit is deficient in this aspect, Black meets the claimed
requirement of “automatically formulating” a search query. HyPursuit is relied
upon for how to categorize documents and restrict searches. It would have been
obvious to implement these strategies in Black to enhance “retrieval of only
relevant information” (p. 180, col. 2) and “specificity of the information need” (id.
at p. 181, col. 1). In any event, it is routine to automate a manual process.
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

21

Leapfrog Enterprises Inc. v. Fisher-Price Inc., 485 F.3d 1157, 1161 (Fed. Cir.
2007).
In sum, for the foregoing reasons and those of the Requester as adopted by
the Examiner, we affirm the rejection of claim 1 as obvious in view of Black and
HyPursuit. Claims 2, 4-7, and 10-20 have substantially the same limitations as
claim 1, and were not argued separately, and therefore fall with claim 1. 37 C.F.R.
§ 41.37(c)(1).
6. BLACK, HYPURSUIT, AND SYSKILL
Claims 3 and 8 stand rejected under 35 U.S.C. § 103(a) as obvious in view
of Black, HyPursuit, and Syskill
10
(RAN 9). This rejection was not separately
argued by Patent Owner. We affirm it for the reasons set forth by the Requester
and as adopted by the Examiner.

7-9. BLACK AND DONALDSON
7. Claims 1, 5, 6, 10-14 and 18 stand rejected under 35 U.S.C. § 103(a) as
obvious in view of Black and Donaldson
11
(RAN 10).
8. Claims 2, 4, 7, 15-17, 19, and 20 stand rejected under 35 U.S.C. § 103(a)
as obvious in view of Black, Donaldson, and Stibel (RAN 10).
Claims 3 and 8 are proposed to be rejected under 35 U.S.C. § 103(a) as
obvious in view of Black, Donaldson, Stibel, and Syskill (RAN 10).
We shall not reach the rejections based on Black and Donaldson as we have
affirmed similar rejections based on Black combined with publications that have
similar disclosures to Donaldson.

10
Pazzani, “Syskill & Webert: Identifying interesting web sites” AAA1 (1996)
(hereinafter “Syskill”).
11
Donaldson, US 7,225,180 B2, issued May 29, 2007.
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

22

10-11. WIESER AND STIBEL
10. Claims 2, 7, 15, and 19 stand rejected under 35 U.S.C. § 103(a) as
obvious in view of Wieser and Stibel
12
(RAN 7).
11. Claims 3 and 8 stand rejected under 35 U.S.C. 103(a) as obvious in view
of Wieser, Stibel, and Syskill (RAN 7).
Patent Owner did not address either of the rejections.
Claims 2, 3, 7, 8, 15, and 19 are dependent upon claims whose rejection over
Wieser alone has been affirmed. We shall affirm the rejections for the reasons set
forth by the Requester and Examiner.

12 and 13. RHODES AND STIBEL OR SYSKILL
12. Claims 4, 16-17 and 20 stand rejected under 35 U.S.C. § 103(a) as
obvious in view of Rhodes in view of Stibel (RAN 8).
13. Claims 3 and 8 stand rejected under 35 U.S.C. § 103(a) as obvious in
view of Rhodes in view of Syskill (RAN 8).
The rejection over Rhodes alone was reversed. Stibel and Syskill are not
said by the Requester or Examiner to make up for the deficiencies of Rhodes. We
are therefore compelled to reverse the rejection of these claims.

14. NEW GROUND OF REJECTION OF CLAIM 9
Claim 9 stands rejected under 35 U.S.C. § 103(a) as obvious in view of
Rhodes in view of Black, Syskill, and Mauldin
13
(RAN 8). Claim 9 depends on

12
Stibel, US 7,089,236 B1, issued August 8, 2006.
13
Mauldin, 5,748,954, issued May 5, 1998.
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

23

claim 1, through intervening claims 8, 7, and 6. Claim 9 pertains to abbreviating
terms extracted by a text categorizer, and comprises a series of recited steps.
The anticipation rejection of claim 1 over Rhodes was reversed because we
found that a preponderance of the evidence did not support a finding that Rhodes
describes steps [3] and [4] of the claim. Although we found that Black describes
these steps, we also noted that Black does not describe restricting a search as
recited in step [3], but found that this deficiency was met by HyPursuit.
Requester provided fact-based reasoning as to why all the limitations of
claim 9 are met by Black, Syskill, and Mauldin (Request 96-98). Requester also
provided fact-based reasoning as to why all the limitations of intervening claims 6
and 7 are met by Black and HyPursuit (id. at 111-112) and claim 8 by Black,
Hypursuit, and Syskill (id. at 121). Patent Owner has not provided arguments or
evidence as to why the limitations in these claims are non-obvious, and we find no
defect in the facts or reasoning set forth by Requester. Consequently, we shall set
forth a new ground of rejection of claim 9 as obvious in view of Rhodes, Black,
HyPursuit, Syskill, and Mauldin.
This decision contains new grounds of rejection pursuant to 37 C.F.R.
§ 41.77(b) which provides that “[a]ny decision which includes a new ground of
rejection pursuant to this paragraph shall not be considered final for judicial
review.” Accordingly, no portion of the decision is final for purposes of judicial
review. A requester may also request rehearing under 37 C.F.R. § 41.79, if
appropriate, however, the Board may elect to defer issuing any decision on such
request for rehearing until such time that a final decision on appeal has been issued
by the Board.
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

24

For further guidance on new grounds of rejection, see 37 C.F.R. § 41.77(b)-
(g). The decision may become final after it has returned to the Board. 37 C.F.R.
§ 41.77(f).
37 C.F.R. § 41.77(b) also provides that the Patent Owner, WITHIN ONE
MONTH FROM THE DATE OF THE DECISION, must exercise one of the
following two options with respect to the new grounds of rejection to avoid
termination of the appeal as to the rejected claims:
(1) Reopen prosecution. The owner may file a response requesting
reopening of prosecution before the examiner. Such a response must be either an
amendment of the claims so rejected or new evidence relating to the claims so
rejected, or both.
(2) Request rehearing. The owner may request that the proceeding be
reheard under § 41.79 by the Board upon the same record. …
Any request to reopen prosecution before the examiner under 37 C.F.R.
§ 41.77(b)(1) shall be limited in scope to the “claims so rejected.” Accordingly, a
request to reopen prosecution is limited to issues raised by the new ground(s) of
rejection entered by the Board. A request to reopen prosecution that includes
issues other than those raised by the new ground(s) is unlikely to be granted.
Furthermore, should the patent owner seek to substitute claims, there is a
presumption that only one substitute claim would be needed to replace a cancelled
claim.
A requester may file comments in reply to a patent owner response.
37 C.F.R. § 41.77(c). Requester comments under 37 C.F.R. § 41.77(c) shall be
limited in scope to the issues raised by the Board's opinion reflecting its decision
to reject the claims and the patent owner's response under paragraph 37 C.F.R.
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

25

§ 41.77(b)(1). A newly proposed rejection is not permitted as a matter of right. A
newly proposed rejection may be appropriate if it is presented to address an
amendment and/or new evidence properly submitted by the patent owner, and is
presented with a brief explanation as to why the newly proposed rejection is now
necessary and why it could not have been presented earlier.
Compliance with the page limits pursuant to 37 C.F.R. § 1.943(b), for all
patent owner responses and requester comments, is required.
The examiner, after the Board’s entry of a patent owner response and
requester comments, will issue a determination under 37 C.F.R. § 41.77(d) as to
whether the Board’s rejection is maintained or has been overcome. The
proceeding will then be returned to the Board together with any comments and
reply submitted by the owner and/or requester under 37 C.F.R. § 41.77(e) for
reconsideration and issuance of a new decision by the Board as provided by
37 C.F.R. § 41.77(f).

AFFIRMED; 41.77(b)
Appeal 2014-000031
Reexamination Control 95/001,430
Patent 6,778,979 B2

26

CC:

Patent Owner:

QUINN EMANUAL URQUHART
& SULLIVAN, LLP
50 California Street, 22
nd
Floor
San Francisco, CA 94111

Third Party Requester:

SCHWEGMAN LUNDBERG & WOESSNER/XEROX
P.O. Box 2938
Minneapolis, MN 55402
lb