Ex Parte Brill et al

Patent Trial and Appeal Board

Mar 26, 2013

10460955 (P.T.A.B. Mar. 26, 2013)

UNITED STATES PATENT AND TRADEMARK OFFICE
________________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
________________

Ex parte ERIC D. BRILL and SOUMYA RAY
________________

Appeal 2010-003575
Application 10/460,955
Technology Center 2100
________________

Before JOSEPH F. RUGGIERO, ERIC B. CHEN, and JOHN A. EVANS,
Administrative Patent Judges.

CHEN, Administrative Patent Judge.

DECISION ON APPEAL

Appeal 2010-003575
Application 10/460,955

2
This is an appeal under 35 U.S.C. § 134(a) from the final rejection of
claims 1-51, all the claims pending in the application. We have jurisdiction
under 35 U.S.C. § 6(b). We affirm-in-part and enter a new ground of
rejection pursuant to our authority under 37 C.F.R. § 41.50(b).

STATEMENT OF CASE
Appellants’ invention relates to an architecture for generating
responses to search engine queries, including question and answer pairs used
to provide a predefined question hierarchy to facilitate a search. Informative
terms and one or more questions are associated with each question node of
the hierarchy and a question posed is mapped into the hierarchy to the most
similar question class node. Weighted informative terms associated with
that node are then returned, and can be used by an automatic question
answering system to more effectively find an answer to a question.
(Abstract.)
Independent claims 1 and 41 are exemplary, with disputed limitations
in italics:
1. A system that facilitates generating responses to queries,
comprising:
a component that receives a query; and
an abstraction component that maps the query into an
abstraction hierarchy according to a distribution of terms found in
potential answer passages for the query, to facilitate identifying an
answer term to the query.

41. A method of generating an abstraction hierarchy,
comprising:
receiving an existing abstraction hierarchy of nodes of
questions and corresponding informative terms;
Appeal 2010-003575
Application 10/460,955

3
receiving a new question;
generating informative terms for the new question;
calculating a similarity value between the new question and
every question in the existing abstraction hierarchy; and
grouping the new question with the most similar question in the
existing abstraction hierarchy based on the similarity value.

Claims 1-12, 17-38, and 44-51 stand rejected under 35 U.S.C.
§ 102(e) as being anticipated by Arnold (U.S. Patent No. 6,745,161 B1;
June 1, 2004; filed July 10, 2000).
1

Claims 13-16 stand rejected under 35 U.S.C. § 103(a) as being
obvious over Arnold and Tamayo (U.S. Patent No. 6,836,773 B2; Dec. 28,
2004; filed Sept. 27, 2001).
Claims 39-43 stand rejected under 35 U.S.C. § 103(a) as being
obvious over Arnold and Lange (U.S. Patent No. 6,236,958 B1; May 22,
2001).

ANALYSIS
§ 102 Rejection – Arnold
First, we are unpersuaded by Appellants’ arguments (App. Br. 5-6;
see also Reply Br. 5) that Arnold does not describe the limitation
“abstraction component,” as recited in independent claim 1.
The Examiner found that the Information Extraction (IE) engine 702
of Arnold corresponds to the claimed “abstraction component.” (Ans. 3;
Arnold, col. 8, ll. 49-55.) We agree with the Examiner.

1
Application has a filing date of June 13, 2003 and Arnold has an issue date
of June 1, 2004. Accordingly, in the statement of the rejection, the
Examiner inadvertently rejected the claims under 35 U.S.C § 102(b).
Appeal 2010-003575
Application 10/460,955

4
One relevant plain meaning of “abstraction” in the context of object-
oriented programming is “the process of reducing an object to its essence so
that only the necessary elements are represented.” MICROSOFT COMPUTER
DICTIONARY 15 (2d ed. 2002). Although Appellants’ Specification provides
no express definition for “abstraction,” the Specification discloses that:
For example, if a user presents the query “Who is Bill Gates
married to”, a text passage that contains the words “Bill”,
“Gates” and “married” is a good candidate passage for
containing an answer. An answer passage that contains the
string “Bill Gates is married to” is an even better candidate
passage. Conventional techniques can handle both of these
cases. However, there are other non-straightforward answer text
patterns, such as “ . . . have three children and live in Redmond,
Wash.”, which is also a good candidate passage, but which
conventional systems cannot effectively process.
(Spec. 6:11-18.) Accordingly, in light of Appellants’ Specification and
applying a plain meaning of “abstraction” in the context of computing, we
interpret “abstraction” as the process of reducing a search query into only
necessary terms (e.g., the query “Who is Bill Gates married to” is reduced
into the necessary terms of “Bill”, “Gates”, and “married” (see Spec. 6:12-
13)).
Arnold relates to concept-based Internet searching (col. 1, ll. 11-14)
that features linguistic pattern recognition of information (Abstract). In one
embodiment, Figure 7 of Arnold illustrates a process of preparing an
information cache (col. 4, ll. 14-16), which includes an Information
Extraction (IE) engine 702, a business news article 704, extracted
concepts 706, and an information frame 708 (col. 8, ll. 56-65). “The
information cache is based on ‘concept-based storage’: that is, the IE engine
identifies and indexes information in a topic-specific normalized format.”
Appeal 2010-003575
Application 10/460,955

5
(Col. 8, ll. 50-52.) Because the IE engine 702 of Arnold reduces a search
query into only necessary terms, Arnold describes an “abstraction
component.”
Appellants argue that the “Final Action incorrectly contends that the
cited document [Arnold] discloses the claimed abstraction component. . . .”
(App. Br. 5; see also Reply Br. 5.) However, contrary to Appellants’
arguments, the IE engine 702 of Arnold corresponds to the claimed
“abstraction component.”
Therefore, we agree with the Examiner that Arnold describes the
limitation “abstraction component.”
Second, we are unpersuaded by Appellants’ arguments (App. Br. 6;
see also Reply Br. 6) that Arnold does not describe the limitation
“abstraction hierarchy,” as recited in independent claim 1.
The Examiner found that the Query Engine 804 of Arnold
corresponds to the claimed “abstraction hierarchy.” (Ans. 3, 16-17; Arnold,
col. 9, ll. 35-45, col. 10, ll. 10-22, figs. 8 and 9.) We agree with the
Examiner.
As discussed previously, one relevant plain meaning of “abstraction”
in the context of object-oriented programming is “the process of reducing an
object to its essence so that only the necessary elements are represented.”
MICROSOFT COMPUTER DICTIONARY 15 (2d ed. 2002). Furthermore, a
relevant definition of “hierarchy” is “[a] type of organization that, like a tree,
branches into more specific units, each of which is ‘owned’ by the higher-
level unit immediately above.” Id. at 252. Although Appellants’
Specification provides no express definition for “abstraction hierarchy,” the
Specification discloses that:
Appeal 2010-003575
Application 10/460,955

6
At 200, question classes are derived in preparation for creating
an abstraction hierarchy for each question class. Such classes
include, Who, What, Where, When, How-many, etc. At 202, an
abstraction hierarchy is constructed for each question class.
Each class hierarchy is further structured with a number of
different nodes according to clusters of informative terms
derived from question/answer pairs.
(Spec. 8:12-16 (emphasis added).) Accordingly, in light of Appellants’
Specification and applying a plain meaning of “abstraction” and “hierarchy”
in the context of computing, we interpret “abstraction hierarchy” as the
process of reducing a search query into only necessary terms and
organization of such necessary terms into braches of more specific units
(e.g., question classes are branched into “Who, What, Where, When, How-
many, etc.” (see Spec. 8:12-14)).
Figure 9 of Arnold illustrates a process flowchart for allowing
concept based information searching. (Col. 4, ll. 20-22.) In reference to
Figure 9, Arnold explains that “[i]n operation 904, the textual information is
parsed to create topic specific information packets which are stored in an
information cache in operation 906.” (Col. 10, ll. 14-16.) Thus, because the
textual information is parsed (i.e., abstracted) and such textual information is
organized and branched according to topic (i.e., a hierarchy), Arnold
describes the limitation “abstraction hierarchy.”
Appellants argue that “an incorrect generalization is made of the
abstraction hierarchy, as claimed and disclosed e.g. at page 8, line 12 et seq.
of the subject disclosure.” (App. Br. 6; see also Reply Br. 6.) However,
when the claimed “abstraction hierarchy” is construed in light of the
Specification, this claim term is broad enough to encompass the
operation 904 of Arnold, in which “the textual information is parsed to
Appeal 2010-003575
Application 10/460,955

7
create topic specific information packets which are stored in an information
cache in operation 906” (col. 10, ll. 14-16).
Therefore, we agree with the Examiner that Arnold describes the
limitation “abstraction hierarchy.”
Third, we are unpersuaded by Appellants’ arguments (App. Br. 6; see
also Reply Br. 6) that Arnold does not describe the limitation “maps the
query into an abstraction hierarchy,” as recited in independent claim 1.
The Examiner found the search constraint set 806 of Figure 8
corresponds to the limitation “maps the query into an abstraction hierarchy.”
(Ans. 16; Arnold, fig. 8.) We agree with the Examiner.
Figure 8 of Arnold illustrates a process for network-based information
retrieval (col. 4, ll. 17-19), including a user’s question 802, a Query
Engine 804, and a search constraint set 806 (col. 9, ll. 35-46). Figure 8 of
Arnold further illustrates that in response to a query “What company did
IBM buy in February?” in the user’s question 802, the search constraint
set 806 provides “Agent: IBM Corp.” and “Date: during Feb 2000.”
Because of the correspondence between “IBM” from the user’s question 802
and “Agent: IBM Corp.” from the search constraint set 806, Arnold
describes the limitation “maps the query into an abstraction hierarchy.”
Appellants argue that “Arnold et al. requires considerable pre-
processing of stored data so as to implement the simple ‘parse and search’
operation” and accordingly, “[t]his is not analogous to the claimed
mapping.” (App. Br. 6; see also Reply Br. 6.) Contrary to Appellants’
arguments, Figure 8 of Arnold illustrates the mapping of terms for the user’s
question 802 to the search constraint set 806.
Appeal 2010-003575
Application 10/460,955

8
Therefore, we agree with the Examiner that Arnold describes the
limitation “maps the query into an abstraction hierarchy.”
Last, we are unpersuaded by Appellants’ arguments (App. Br. 5; see
also Reply Br. 6) that Arnold does not describe the limitation “according to
a distribution of terms found in potential answer passages for the query,” as
recited in independent claim 1.
The Examiner found that the retrieval of information from an
information cache of Arnold corresponds to the limitation “according to a
distribution of terms found in potential answer passages for the query.”
(Ans. 3; Arnold, col. 8, ll. 49-55.) We agree with the Examiner.
As discussed previously, Figure 7 of Arnold illustrates a process of
preparing an information cache. (Col. 4, ll. 14-16.) Furthermore, this
information cache is based on “the IE engine [that] identifies and indexes
information in a topic-specific normalized format” which “enables
information retrieval based on concept, independently from the way the
information may have been expressed in the source documents or in the
user’s query.” (Col. 8, ll. 49-55.) Thus, Arnold describes the limitation
“according to a distribution of terms found in potential answer passages for
the query.”
Appellants argue that “[n]either does the cited document [Arnold]
disclose anything analogous to the claimed abstraction hierarchy, which
utilizes potential answer passages having a distribution of terms, as
claimed.” (App. Br. 6; see also Reply Br. 6.) However, as described
previously, the information cache of Arnold that “enables information
retrieval based on concept” (col. 8, l. 53) corresponds to the claim limitation
Appeal 2010-003575
Application 10/460,955

9
“according to a distribution of terms found in potential answer passages for
the query.”
Therefore, we agree with the Examiner that Arnold describes the
limitation “according to a distribution of terms found in potential answer
passages for the query.”
Accordingly, we sustain the rejection of independent claim 1 under 35
U.S.C. § 102(e). Claims 2-12 and 17-19 depend from claim 1, and
Appellants have not presented any substantive arguments with respect to
these claims. Therefore, we sustain the rejection of claims 2-12 and 17-19
under 35 U.S.C. § 102(e), for the same reasons discussed with respect to
independent claim 1.
Independent claims 20, 25, 38, 44 recite limitations similar to those
discussed with respect to independent claim 1, and Appellants have not
presented any substantive arguments with respect to these claims. We
sustain the rejection of claims 20, 25, 38, 44, as well as dependent
claims 21-24, 26-37, and 45-51, for the same reasons discussed with respect
to claim 1.

§ 103 Rejection – Arnold and Tamayo
Although Appellants nominally argue the rejection of dependent
claims 13-16 separately (App. Br. 7), the arguments presented do not point
out with particularity or explain why the limitations of the dependent claims
are separately patentable. Instead, Appellants summarily allege that “Arnold
et al. does not disclose or suggest every limitation set forth in the subject
independent claim, and Tamayo et al. does not cure the aforementioned
deficiencies.” (Id.) We are not persuaded by these arguments for the
Appeal 2010-003575
Application 10/460,955

10
reasons discussed with respect to claim 1, from which claims 13-16 depend.
Accordingly, we sustain this rejection.

§ 103 Rejection – Arnold and Tamayo
Claims 39 and 40
Although Appellants nominally argue the rejection of dependent
claims 39 and 40 separately (App. Br. 7-8), the arguments presented do not
point out with particularity or explain why the limitations of the dependent
claims are separately patentable. Instead, Appellants summarily allege that
“Arnold et al. does not disclose or suggest every limitation set forth in the
subject independent claim, and Lange et al. does not cure the
aforementioned deficiencies.” (Id.) We are not persuaded by these
arguments for the reasons discussed with respect to claim 38, from which
claims 39 and 40 depend. Accordingly, we sustain this rejection.

Claims 41-43
We are persuaded by Appellants’ arguments (App. Br. 8-9; see also
Reply Br. 8-10) the combination of Arnold and Lange would not have
rendered obvious independent claim 41, which includes the limitations
“calculating a similarity value between the new question and every question
in the existing abstraction hierarchy” and “grouping the new question with
the most similar question in the existing abstraction hierarchy based on the
similarity value.”
The Examiner acknowledged that Arnold does not disclose the
limitations “calculating a similarity value between the new question and
every question in the existing abstraction hierarchy” and “grouping the new
Appeal 2010-003575
Application 10/460,955

11
question with the most similar question in the existing abstraction hierarchy
based on the similarity value,” (Ans. 12-13) and thus, relied upon Lange for
teaching a terminology extraction system, in which a source text is aligned
with a target text (Ans. 13-14). The Examiner concluded that “[i]t would
have been obvious . . . to combine the teachings of the cited references
because teaching Lange’s would have allowed Arnold’s automatically
extract information.” (Ans. 14.) We do not agree.
Lange relates to machine translation systems (col. 1, l. 8) for
automatic creation of bilingual terminology (Abstract), such that sequences
of source terms are aligned with sequences of target terms (col. 3, ll. 22-23).
The method of Lange includes “computing an association score for each pair
of linked source and target nodes” (col. 3, ll. 36-37) (i.e., the claimed
“calculating a similarity value”) followed by “selection of relevant bilingual
terms is made according to the result of these alignment statistics” (col. 3,
ll. 47-49) (i.e., the claimed “grouping the new question with the most similar
question”).
Claim 41 recites “calculating a similarity value between the new
question and every question in the existing abstraction hierarchy” and
“grouping the new question with the most similar question in the existing
abstraction hierarchy . . .” (emphasis added). Although the terminology
extraction system of Lange aligns a source text (i.e., the claimed “new
question”) with a target text (i.e., the claimed “similar question”) and
calculates an association score, the Examiner has provided insufficient
evidence to support a finding that Lange teaches the limitation “in the
existing abstraction hierarchy.”
Appeal 2010-003575
Application 10/460,955

12
Thus, we do not agree with the Examiner that the combination of
Arnold and Lange would have rendered obvious independent claim 41,
which includes the limitations “calculating a similarity value between the
new question and every question in the existing abstraction hierarchy” and
“grouping the new question with the most similar question in the existing
abstraction hierarchy based on the similarity value.”
Accordingly, we do not sustain the rejection of independent claim 41
under 35 U.S.C. § 103(a). Claims 42 and 43 depend from independent
claim 41. We do not sustain the rejection of claims 42 and 43 under 35
U.S.C. § 103(a) for the same reasons discussed with respect to independent
claim 41.

NEW GROUND OF REJECTION UNDER 37 C.F.R. § 41.50(b)
We enter the following new ground of rejection:
Claims 25-43 are rejected under 35 U.S.C. § 101 as being directed to
non-statutory subject matter.
“Phenomena of nature, though just discovered, mental processes, and
abstract intellectual concepts are not patentable, as they are the basic tools of
scientific and technological work.” Gottschalk v. Benson, 409 U.S. 63, 67
(1972) (emphasis added). Independent claims 25, 38, and 41 do not
expressly require the method to be performed by any machine, nor do the
claims require a transformation of a physical object or data. In particular,
the steps recited in claim 25, which include receiving a query, processing the
query for informative terms, and identifying an answer term to the query,
can be performed in the human mind. Claims 38 and 41 recite similar
limitations, which can also be performed in the human mind.
Appeal 2010-003575
Application 10/460,955

13
This decision contains a new ground of rejection pursuant to
37 C.F.R. § 41.50(b).
37 C.F.R. § 41.50(b) provides that a “new ground of rejection
pursuant to this paragraph shall not be considered final for judicial review.”
37 C.F.R. § 41.50(b) also provides that Appellants, WITHIN TWO
MONTHS FROM THE DATE OF THE DECISION, must exercise one of
the following two options with respect to the new ground of rejection to
avoid termination of proceedings (37 C.F.R. § 1.197 (b)) as to the rejected
claims:
(1) Reopen prosecution. Submit an appropriate
amendment of the claims so rejected or new evidence relating
to the claims so rejected, or both, and have the matter
reconsidered by the examiner, in which event the proceeding
will be remanded to the examiner ….
(2) Request rehearing. Request that the proceeding be
reheard under 37 C.F.R. § 41.52 by the Board upon the same
record ….

DECISION
The Examiner’s decision to reject claims 1-12, 17-38, and 44-51
under 35 U.S.C. § 102(e) is affirmed.
The Examiner’s decision to reject claims 13-16, 39, and 40 under 35
U.S.C. § 103(a) is affirmed.
The Examiner’s decision to reject claims 41-43 under 35 U.S.C.
§ 103(a) is reversed.
New ground of rejection has been entered under 37 C.F.R. § 41.50(b)
for claims 25-43, rejected under 35 U.S.C. § 101 as being directed to
non-statutory subject matter.
Appeal 2010-003575
Application 10/460,955

14
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a)(1)(iv).

AFFIRMED-IN-PART
37 C.F.R. § 41.50(b)

ke