`571-272-7822
`
`
`
`
`
`
`
`
` Paper No. 9
`
`
` Entered: September 6, 2016
`
`
`
`
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`____________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________
`
`GOOGLE INC.,
`Petitioner,
`
`v.
`
`IMRPOVED SEARCH LLC,
`Patent Owner.
`____________
`
`Case IPR2016-00797
`Patent 6,604,101 B1
`____________
`
`
`
`Before JONI Y. CHANG, PATRIC R. SCANLON, and
`JACQUELINE T. HARLOW, Administrative Patent Judges.
`
`HARLOW, Administrative Patent Judge.
`
`
`DECISION
`Denying Institution of Inter Partes Review
`37 C.F.R. § 42.108
`
`
`
`
`
`
`
`
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`I.
`
`INTRODUCTION
`
`Petitioner, Google Inc. (“Google”), filed a Petition requesting an inter
`
`partes review of claims 1, 2, 4–7, 12, 13, 15–17, and 22–28 of U.S. Patent
`
`No. 6,604,101 B1 (Ex. 1001, “the ’101 patent”). Paper 1 (“Pet.”). Patent
`
`Owner, Improved Search LLC, (“Improved Search”), filed a Preliminary
`
`Response. Paper 8 (“Prelim. Resp.”). We have jurisdiction under 35 U.S.C.
`
`§ 314, which provides that an inter partes review may not be instituted
`
`unless the information presented in the petition “shows that there is a
`
`reasonable likelihood that the petitioner would prevail with respect to at
`
`least 1 of the claims challenged in the petition.”
`
`For the reasons set forth below, we deny the Petition.
`
`A. Related Matters
`
`The ’101 patent is asserted against AOL Inc., identified by Google as
`
`a real party-in-interest to the Petition (Pet. 2), in Improved Search LLC v.
`
`AOL Inc., Civ. No. 15-262 (D. Del.). Id. at 3.
`
`B. The ’101 Patent
`
`The ’101 patent is titled “Method and System for Translingual
`
`Translation of Query and Search and Retrieval of Multilingual Information
`
`on a Computer Network.” Ex. 1001, [54]. The ’101 patent issued from U.S.
`
`Patent Application No. 09/606,655, filed on June 28, 2000. Id. [21], [22].
`
`The ’101 patent describes “a method and system for conducting a
`
`translingual search on the Internet and accessing multilingual web sites
`
`through dialectal standardization, pre-search translation and post-search
`
`2
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`translation.” Id. at 1:11–14. More specifically, the ’101 patent discloses a
`
`system and method “for dialectally standardizing a query input by the user in
`
`the source language and then translating the standardized keyword to the
`
`target language and searching and retrieving web documents in the target
`
`language as well as providing translations of said search results into the
`
`source language.” Id. at 3:40–45.
`
`In discussing the problem it aims to solve, the ’101 patent notes that
`
`“most of the search tools cater primarily to the needs of the English speaking
`
`Internet user,” and explains that “[t]his is a serious drawback, which has not
`
`been addressed by any of the existing search engines.” Id. at 2:14–24. The
`
`’101 patent likewise observes that “the non-English speaking Internet users
`
`also create web sites to store information in non-English languages. This
`
`rich source of information is not available to query by English oriented
`
`search engines.” Id. at 2:25–28.
`
`C. Illustrative Claim
`
`Of the challenged claims, claims 1, 12, 22, and 23 are independent.
`
`Claim 1, reproduced below, is illustrative of the claimed subject matter.
`
`A method for performing a contextual search and
`1.
`retrieval of documents in a computer network, comprising:
`
`receiving through an input device, a query in a first
`language;
`
`processing said query to extract at least one content word
`from the query;
`
`performing dialectal standardization of the at least one
`content word extracted from the query;
`
`3
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`translating the at least one dialectally standardized content
`word into a second language through a translator;
`
`performing a contextual search in the second language
`based on the at least one translated content word, using a search
`engine in the second language; and
`
`obtaining the search results in the second language in the
`form of at least one of site names (URLs) and documents,
`satisfying a search criteria.
`
`Ex. 1001, 7:66–8:15.
`
`D. Prior Art Relied Upon
`
`Google relies upon the following prior art references (Pet. 4–5):
`
`Fluhr et al., Multilingual Database and Crosslingual Interrogation in a Real
`Internet Application: Architecture and Problems of Implementation, Cross-
`Language Text & Speech Retrieval: Papers from the 1997 AAAI Spring
`Symposium, Technical Report SS-97-05, (AAAI Press) 32 (1997)
`(“Fluhr ’97) (Ex. 1003).
`
`Fluhr et al., Distributed Cross-Lingual Information Retrieval,” Cross-
`Language Information Retrieval, 41 (Gregory Grefenstette ed., 1998)
`(“Fluhr ’98) (Ex. 1004).
`
`Yamabana et al., A Language Conversion Front-End for Cross-Language
`Information Retrieval, Cross-Language Information Retrieval, 93 (Gregory
`Grefenstette ed., 1998) (“Yamabana”) (Ex. 1005).
`
`Bian, Integrating Query Translation and Document Translation in a Cross-
`Language Information Retrieval System, Machine Translation and the
`Information Soup: Third Conference of the Association for Machine
`Translation in the Americas, AMTA’98, Langhorne, PA, USA, October 28–
`31, 1998 Proceedings, 250 (David Farwell et al. eds., Springer-Verlag Berlin
`Heidelberg 1998) (“Bian”) (Ex. 1006).
`
`4
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`E. Asserted Grounds of Unpatentability
`
`Google asserts the following grounds of unpatentability (Pet. 5–6):
`
`Claim(s)
`
`Basis
`
`Reference(s)
`
`1, 2, 4, 5, 22, 24,
`27, 28
`6, 7, 23, 25
`28
`12, 13, 15, 16, 26
`
`17
`
`§ 103(a) Fluhr ’97 and Fluhr ’98
`
`§ 103(a) Fluhr ’97, Fluhr ’98, and Yamabana
`§ 103(a) Fluhr ’97, Fluhr ’98, and Bian
`§ 103(a) Bian, Fluhr ’97, and Fluhr ’98
`Bian, Fluhr ’97, Fluhr ’98, and
`Yamabana
`
`§ 103(a)
`
`II. ANALYSIS
`
`A. Claim Construction
`
`In an inter partes review, claim terms in an unexpired patent are given
`
`their broadest reasonable interpretation in light of the specification of the
`
`patent in which they appear. 37 C.F.R. § 42.100(b). Under the broadest
`
`reasonable interpretation standard, claim terms are given their ordinary and
`
`customary meaning as would be understood by one of ordinary skill in the
`
`art in the context of the entire disclosure. In re Translogic Tech., Inc.,
`
`504 F.3d 1249, 1257 (Fed. Cir. 2007).
`
`Even under this standard, “the Board’s construction cannot be
`
`divorced from the specification and the record evidence, . . . and must be
`
`consistent with the one that those skilled in the art would reach.” Microsoft
`
`Corp. v. Proxyconn, Inc., 789 F.3d 1292, 1298 (Fed. Cir. 2015) (noting that
`
`the Board may not “construe claims during IPR so broadly that is
`
`constructions are unreasonable under general claim construction principles”)
`
`5
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`(internal quotations and citations omitted). The specification remains “the
`
`single best guide to the meaning of a disputed term and . . . acts as a
`
`dictionary when it expressly defines terms used in the claims or when it
`
`defines terms by implication.” SightSound Techs., LLC v. Apple Inc., 809
`
`F.3d 1307, 1317 (Fed. Cir. 2015) (internal quotation omitted). “Thus a
`
`claim term may be clearly redefined without an explicit statement of
`
`redefinition.” Id. (internal quotation omitted).
`
`Only those terms that are in controversy need be construed, and only
`
`to the extent necessary to resolve the controversy. Vivid Techs., Inc. v. Am.
`
`Sci. & Eng’g, Inc., 200 F.3d 795, 803 (Fed. Cir. 1999).
`
`“contextual search” / “contextually searching”
`
`Each challenged independent claim (claims 1, 12, 22, and 23) recites
`
`the term “contextual search” or “contextually searching.” Ex. 1001, 7:66–
`
`8:15, 9:4–19, 10:15–39.
`
`Google asserts that the term “contextual search” should be construed
`
`to mean “identification of one or more documents from a larger collection
`
`based on the presence of specific words contained in the text of those
`
`documents.” Pet. 12–13. Google observes that the term “contextual search”
`
`does not appear in the Specification, but notes that “the [S]pecification
`
`describes using search engines to find documents based on words in the text
`
`of those documents.” Id. Google argues also that
`
`The prosecution history makes clear that the term “contextual
`search” was included to distinguish a search to identify
`documents based on the presence of words in the text of the
`documents from a search to locate a known document with
`
`6
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`particular characteristics (e.g., the latest version of a particular
`paper), because the latter type of search was present in a prior art
`reference cited by the Examiner.
`
`
`
`Id.
`
`Improved Search counters that “contextual search” means
`
`“identification of relevant documents over the Web or computer network
`
`based on words contained in the documents.” Prelim. Resp. 23. Improved
`
`Search further clarifies that under its proposed construction, “the search
`
`extends to an external set of unidentified documents over a computer
`
`network or the Web, and not a local, pre-identified set of documents as in a
`
`document database.” Id.
`
`In support of its position, Improved Search observes that “[t]he
`
`[S]pecification recites not only the need to search for documents on a
`
`computer network such as the Internet, but specifically describes the
`
`invention in terms of searching Web documents.” Id. at 24. Improved
`
`Search additionally argues that Google’s proposed construction is
`
`inconsistent with the plain language of claims 1 and 12, which recite
`
`“obtaining the search results in the second language in the form of at least
`
`one of site names (URLs) and documents, satisfying a search criteria”
`
`(Ex. 1001, 7:66–8:15), and “a search engine for searching in the second
`
`language, the site names (URLs), pages and descriptions satisfying search
`
`criteria” (id. at 9:4–19).
`
`Improved Search also avers that the Applicant amended independent
`
`claims 1, 12, 22, and 23 to recite “contextual search” to “further distinguish
`
`a feature of the present invention allowing . . . a contextual search of the
`
`7
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`Internet in a second or target language” (Ex. 2002, 106). Prelim. Resp. 23.
`
`Improved Search further contends that Dr. Oard’s testimony that in
`
`contextual searching, “the user only provides a query including some terms
`
`of interest that may or may not appear in the unidentified documents”
`
`(Ex. 1002 ¶ 80) supports a construction of “contextual search” requiring that
`
`“the search extends to an external set of unidentified documents over a
`
`computer network or the Web, and not a local, pre-identified set of
`
`documents as in a document database.” Prelim. Resp. 23.
`
`On this record, and for purposes of this Decision, in view of the
`
`repeated, consistent, and exclusive teachings of the Specification, as well as
`
`the language of the claims themselves, and the prosecution history, we
`
`interpret the claim terms “contextual search” and “contextually searching” to
`
`mean “identification of/identifying relevant documents from the domain-
`
`unlimited set of documents available on the World Wide Web, based on
`
`words contained in the documents.”
`
`The disclosure of the ’101 patent is expressly addressed to the
`
`problem of, and a solution for, searching the vast array of multilingual
`
`websites available via the Internet. The Field of Invention explains that the
`
`invention is specifically directed to multilingual web searching: “[t]his
`
`invention relates generally to translation of query and retrieval of
`
`multilingual information on the web and more particularly to a method and
`
`system for conducting a translingual search on the Internet and accessing
`
`multilingual web sites through dialectal standardization, pre-search
`
`translation and post-search translation.” Ex. 1001, 1:9–14.
`
`8
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`The Description of the Prior Art focuses on “the sheer volume of
`
`documents available on different sites on the World Wide Web (‘Web’)” as
`
`necessitating the development of advanced search tools (id. at 1:16–20), and
`
`directs particular focus toward the shortcomings of then-existing web search
`
`tools with regard to multilingual search (id. at 2:9–3:23). The ’101 patent
`
`repeatedly identifies the need for a system to help “users to transcend
`
`language barriers while making a search on the web.” Id. at 2:61–63; see
`
`also 2:9–60, 2:63–3:23.
`
`More to the point, the Specification “repeatedly, consistently, and
`
`exclusively” describes the invention of the ’101 patent as an improved
`
`system and method for web search. In re Abbott Diabetes Care Inc., 696
`
`F.3d 1142, 1150 (Fed. Cir. 2012) (finding that the claim term
`
`“electrochemical sensor” excluded cables and wires based on the repeated,
`
`consistent, and exclusive depiction in the Specification of an electrochemical
`
`sensor without external cables or wires and disparagement of sensors with
`
`external cables or wires) (internal quotation omitted); see also SightSound
`
`Techs., LLC, 809 F.3d at 1317 (Fed. Cir. 2015) (“Thus a claim term may be
`
`clearly redefined without an explicit statement of redefinition.” (internal
`
`quotation omitted)).
`
`For example, the Abstract and Summary of Invention each explain
`
`that the ’101 patent teaches a method for “searching and retrieving web
`
`documents in the target language.” Id. at Abstract, 3:40–45; see also id. at
`
`4:10–16. The Detailed Description of Invention likewise explains that:
`
`9
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`The invention incorporates a new and unique methodology
`and system for translingual translation of query and search and
`retrieval of multilingual web documents. Such a system enables
`a user to access web documents in a target language other than
`his/her own source language with the option of having these web
`documents translated back either in part or in whole into the
`source language.
`
`Id. at 4:61–67; see also id. at 6:17–22, 6:40–44, 6:57–63, 7:28–32. The
`
`Figures of the ’101 patent are similarly described as illustrating systems and
`
`methods for performing search and retrieval of multilingual web documents.
`
`Id. at 5:5–21, 7:1–33.
`
`Although the Specification does not include an explicit statement
`
`disclaiming searches of fixed databases or identified document populations
`
`from claim scope, “this is not an instance where the [S]pecification would
`
`necessarily have to disavow an embodiment that would otherwise be
`
`covered by the plain language of the claims.” In re Abbott Diabetes Care
`
`Inc., 696 F.3d at 1149. Rather, claim terms like “a search engine for
`
`searching . . . the site names (URLs), pages and descriptions satisfying
`
`search criteria” (Ex. 1001, 9:13–15 (claim 12)), “obtaining the search results
`
`. . . in the form of at least one of site names (URLs), pages and descriptions”
`
`(id. at 8:13–15 (claim 1)), and “translating said web documents” (id. at
`
`10:29–30 (claim 23)) are “entirely consistent with and even support the
`
`[S]pecification’s exclusive depiction of” contextual search as a search
`
`performed on an unidentified document set over the web or equivalent
`
`computer network. In re Abbott Diabetes Care Inc., 696 F.3d at 1149.
`
`10
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`Google does not proffer evidence or reasoning sufficient to support a
`
`broader construction of “contextual search” or “contextually searching” that
`
`would encompass a search performed on a fixed database, or on a
`
`pre-selected set of documents. Instead, Google’s observation that “the
`
`[S]pecification describes using search engines to find documents based on
`
`words in the text of those documents” (Pet. 13) is consistent with our
`
`construction, as the ’101 patent differentiates between “web directories,”
`
`which are identified as “database[s] comprising [] selected web sites,” and
`
`“search engines,” which “crawl around the Web and index, and catalogue
`
`the contents from different web sites into the database of the search engine
`
`itself.” Ex. 1001, 1:45–54.
`
`Neither do the portions of the prosecution history referenced by
`
`Google support a broader construction, as those excerpts do not address
`
`whether the claims require web-searching, or extend to searches performed
`
`on fixed document sets. Moreover, we observe that the prosecution history
`
`explicitly states that “[i]ndependent claims 1, 9, 12, 19, 22 and 23 have
`
`herein been amended to further distinguish a feature of the present invention
`
`allowing input of a query in a first language or source language; and a
`
`contextual search of the Internet in a second or target language.” Ex. 2002,
`
`10 (original emphasis omitted, emphasis added).
`
`Accordingly, based on the implicit definition set forth in the
`
`Specification, the claims themselves, and the prosecution history, we
`
`construe “contextual search” and “contextually searching” broadly, but
`
`reasonably, as “identification of/identifying relevant documents from the
`
`11
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`domain-unlimited set of documents available on the World Wide Web,
`
`based on words contained in the documents.”
`
`Other claim terms
`
`Although the parties propose constructions for several additional
`
`claim terms, we determine, for purposes of this Decision, none of these
`
`additional terms requires express construction beyond its ordinary and
`
`customary meaning. See Vivid Techs., Inc., 200 F.3d at 803.
`
`B. Obviousness Grounds of Unpatentability
`Based on Fluhr ’97 and Fluhr ’98
`
`Google asserts that claims 1, 2, 4, 5, 22, 24, 27, and 28 are
`
`unpatentable under § 103(a) as obvious, in view of the combination of
`
`Fluhr ’97 and Fluhr ’98 (collectively, “the Fluhr references”). Pet. 19–35.
`
`Claims 1 and 22 are independent. Claims 2, 4, 5, 24, 27, and 28 depend
`
`directly from claim 1. Google relies upon the Declaration of Dr. Douglas
`
`W. Oard (“Oard Declaration”) (Ex. 1002) to support its positions.
`
`Of particular relevance to our analysis, Improved Search asserts that
`
`the Fluhr references do not render obvious a method for performing a
`
`contextual search in a computer network, as required by claim 1, because the
`
`Fluhr references “rely on a domain-specific document database in order to
`
`filter possible translations of individual words in the query.” Prelim.
`
`Resp. 37. Improved Search argues further that “Web searching is very
`
`different from document retrieval over a fixed collection of pre-analyzed
`
`documents.” Id. at 38.
`
`12
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`Improved Search also contends that prior to the invention of the
`
`’101 patent, a long felt need existed for a “one stop shop” to dialectally
`
`standardize one or more words in a query, and then translate the
`
`standardized query prior to performing a search in a different language, in
`
`order to help users transcend language barriers in performing a web search.
`
`Id. at 41. Improved Search states that the Fluhr references acknowledge the
`
`“need for cross-language information retrieval across the World Wide Web,
`
`but the cited references only responded to this problem by providing
`
`cross-language searching of fixed document databases through an
`
`application available by way of the Internet,” and thus fail to remedy the
`
`underlying problem of “transcending language barriers while Web
`
`searching.” Id. Specifically, Improved Search argues that unlike the fixed
`
`database described by the Fluhr references, in which the database searched
`
`can be used as a filter and for selecting translations prior to search, the Web
`
`cannot be used as a filter to narrow the set of possible translations to those
`
`most relevant to the original query. Id. at 42.
`
`Overview of Fluhr ’97
`
`Fluhr ’97 discloses a method for crosslingual interrogation of full-text
`
`databases. Ex. 1003, 2. In particular, Fluhr discloses a system that
`
`incorporates the European Multilingual Information Retrieval (“EMIR”)
`
`project into the Syntactic and Probabilistic Index and Retrieval of
`
`Information in Texts (“SPIRIT”) system to facilitate translingual query of
`
`documents in fixed, multilingual databases. Id.
`
`13
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`The translation and filtering system described by Fluhr ’97 is
`
`explicitly limited to fixed database applications, and in fact depends on the
`
`fixed character of the databases queried to function. Fluhr ’97 teaches that a
`
`query in a source language is made on a particular database, containing a
`
`fixed set of documents in a target language. Id. at 4. The system next
`
`generates all possible translations of the query from the source language to
`
`the target language. Id. Fluhr ’97 teaches that the query translations are
`
`then filtered using the “database lexicon,” to exclude any translations
`
`“incompatible with the domain” queried. Id. The final translation is
`
`subsequently reached by “filtering by the best document[s] that contain” the
`
`translated terms identified via the database lexicon filtering step. Id.
`
`Fluhr ’97 teaches that by 1995, the SPIRIT server was accessible
`
`using standard Internet clients, and further that the card catalog for Fluhr’s
`
`agency was available online at the time of publication for Fluhr ’97. Id. at 5.
`
`Critically, however, only the fixed SPIRIT databases, and not the World
`
`Wide Web, could be queried using the SPIRIT web interface. Id.
`
`Overview of Fluhr ’98
`
`Fluhr ’98 builds on the research disclosed in Fluhr ’97, and discloses a
`
`system for distributed crosslingual information retrieval. Ex. 1004, 3. Fluhr
`
`’98 provides additional detail regarding the EMIR system, and addresses the
`
`problem of searching a “multilingual database” by “splitting the multilingual
`
`database into as many databases as languages.” Id.
`
`14
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`“performing a contextual search”
`
`Independent claim 1 requires “performing a contextual search in the
`
`second language based on the at least one translated content word.”
`
`Ex. 1001, 8:9–11. Independent claim 22 similarly requires “performing a
`
`contextual search in the second language, based on the content word.” Id. at
`
`10:24–25. As set forth above, based on the record before us, and for
`
`purposes of this Decision, we construe “contextual search” to mean
`
`“identification of relevant documents from the domain-unlimited set of
`
`documents available on the World Wide Web, based on words contained in
`
`the documents.”
`
`Google contends that the Fluhr references render obvious these
`
`limitations of claims 1 and 22. Pet. 27–30; see also Ex. 1002 ¶¶ 108, 141.
`
`Google acknowledges, however, that Fluhr ’97 discloses querying “a
`
`monolingual database” (Ex. 1003, 5), and Fluhr ’98 discloses querying “a
`
`French, English or German database (Ex. 1004, 3). Pet. 27–30; see also
`
`Ex. 1002 ¶¶ 108, 141. Google recognizes also that although SPIRIT was
`
`accessible through a “Web interface,” it nevertheless searched and retrieved
`
`only those documents included in the “collection” contained in the SPIRIT
`
`databases. Pet. 21; see also Ex. 1002 ¶¶ 109, 141.
`
`Moreover, neither Google nor Dr. Oard offers evidence or reasoning
`
`to support the conclusion that the fixed database translingual query system
`
`disclosed by the Fluhr references would render obvious the identification of
`
`relevant documents from the domain-unlimited set of documents available
`
`on the World Wide Web, as required by claims 1 and 22.
`
`15
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`Accordingly, Google has not shown a reasonable likelihood of
`
`prevailing in establishing that it would have been obvious to perform a
`
`contextual search in the second language, based on at least one translated
`
`content word, as required by claims 1 and 22.
`
`Conclusion
`
`For the foregoing reasons, we determine that the information
`
`presented in the Petition fails to establish a reasonable likelihood that
`
`Google would prevail in challenging claims 1, 2, 4, 5, 22, 24, 27, and 28 of
`
`the ’101 patent.
`
`C. Obviousness Grounds of Unpatentability
`Based Fluhr ’97, Fluhr ’98, and Yamabana
`
`Google asserts that claims 6, 7, 23, and 25 are unpatentable under
`
`§ 103(a) as obvious in view of the combination of Fluhr ’97, Fluhr ’98, and
`
`Yamabana. Pet. 36–43. Claims 6, 7, and 25 depend, directly or indirectly,
`
`from claim 1. Claim 23 is independent. Google relies on the Oard
`
`Declaration (Ex. 1002) to support its positions.
`
`Of relevance to our analysis, Improved Search reiterates its position
`
`that the combination of Fluhr ’97 and Fluhr ’98 does not render obvious
`
`independent claim 1, and asserts that the same reasoning applies to claim 23.
`
`Improved Search further asserts that Yamabana fails to cure the deficiencies
`
`of the Fluhr references.
`
`16
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`Overview of Yamabana
`
`Yamabana discloses a “disambiguation method for query translation
`
`in cross-language information retrieval.” Ex. 1005, 3. More specifically,
`
`Yamabana teaches “a lexicon-based query translation method in which a
`
`statistical word selection [DM92, DM93] is combined with an interactive
`
`user interface [MASA94, YDK+95] for disambiguation.” Id. at 4.
`
`“contextually searching”
`
`Independent claim 23 recites, in pertinent part, “contextually
`
`searching and obtaining search results in said second language.” Ex. 1001,
`
`10:37–38. As set forth above, based on the record before us, and for
`
`purposes of this Decision, we construe “contextually searching” to mean
`
`“identifying relevant documents from the domain-unlimited set of
`
`documents available on the World Wide Web, based on words contained in
`
`the documents.”
`
`Google does not set forth independent reasons to support the
`
`conclusion that the “contextually searching and obtaining search results in
`
`said second language” element of claim 23 is obvious, but rather, relies on
`
`the same arguments described above with regard to the “performing a
`
`contextual search in the second language based on the at least one translated
`
`content word” step of claim 1. Pet. 39, 43; Ex. 1002 ¶ 165. Furthermore,
`
`Google does not rely on Yamabana to support its position regarding the
`
`obviousness of either claim 1 or claim 23.
`
`Accordingly, for the reasons set forth above, in Part II.B., we
`
`conclude that Google has not shown a reasonable likelihood of prevailing in
`
`17
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`showing that it would have been obvious to perform a contextual search in
`
`the second language, based on at least one translated content word, as
`
`required by claim 1, or to contextually search and obtain search results in a
`
`second language, as required by claim 23.
`
`Conclusion
`
`For the foregoing reasons, we determine that the information
`
`presented in the Petition fails to establish a reasonable likelihood that
`
`Google would prevail in challenging claims 6, 7, 23, and 25 of the
`
`’101 patent.
`
`D. Obviousness Grounds of Unpatentability
`Based on Fluhr ’97, Fluhr ’98, and Bian
`
`Google asserts that claim 28 is unpatentable under § 103(a) as obvious
`
`in view of the combination of Fluhr ’97, Fluhr ’98, and Bian. Pet. 43–44.
`
`Google relies on the Oard Declaration (Ex. 1002) to support its positions.
`
`Relevant to our analysis, Improved Search reiterates its position that
`
`the Fluhr references do not render obvious claim 1, from which claim 28
`
`depends, and argues further that Bian fails to remedy the deficiencies of the
`
`Fluhr references with respect to claim 1, and that in any event, an ordinarily
`
`skilled artisan would not have been motivated to combine Bian with the
`
`Fluhr references. Prelim. Resp. 46–48.
`
`Overview of Bian
`
`Bian discloses a system that integrates machine translation with
`
`information retrieval, using known search engines. Ex. 1006, 4, 11–12.
`
`18
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`Bian describes a dictionary-based approach to query translation and
`
`information retrieval consisting of three major steps:
`
`1. Word segmentation: To identify the word boundary of the
`input stream of Chinese characters.
`
`2. Query translation: To construct the translated English query
`using the bilingual dictionary. The translation disambiguation is
`done using the monolingual corpus.
`
`3. Monolingual IR: To search the relevant documents using the
`translated queries.
`
`Id.
`
`“performing a contextual search”
`
`Google relies on the same arguments described in Part II.B.,
`
`concerning the obviousness of the “performing a contextual search in the
`
`second language based on the at least one translated content word” step of
`
`claim 1. Pet. 39, 43; Ex. 1002 ¶ 165. Google does not rely on Bian to
`
`support its position regarding the obviousness of claim 1.
`
`Accordingly, for the same reasons set forth above we conclude that
`
`Google has not shown a reasonable likelihood of prevailing in showing that
`
`it would have been obvious to perform a contextual search in the second
`
`language, based on at least one translated content word, as required by
`
`claim 28.
`
`Conclusion
`
`For the foregoing reasons, we determine that the information
`
`presented in the Petition fails to establish a reasonable likelihood that
`
`Google would prevail in challenging claim 28 of the ’101 patent.
`
`19
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`E. Obviousness Grounds of Unpatentability
`Based on Bian, Fluhr ’97, and Fluhr ’98
`
`Google asserts that claims 12, 13, 15, 16, and 26 are unpatentable
`
`under § 103(a) as obvious in view of the combination of Bian, Fluhr ’97, and
`
`Fluhr ’98. Pet. 43–57. Claims 13, 15, 16, and 26 depend directly from
`
`independent claim 12. Google relies upon the Oard Declaration (Ex. 1002)
`
`to support its positions.
`
`Of particular relevance to our analysis, Improved Search asserts that
`
`an ordinarily skilled artisan would not have been motivated to combine Bian
`
`with the Fluhr references because Bian is concerned with integrating
`
`machine translation and information retrieval processes, while the Fluhr
`
`references address the problem of querying a fixed document database in a
`
`language that is different from the database language. Prelim. Resp. 48.
`
`Rationale to Combine
`
`Google contends that an ordinarily skilled artisan would have been
`
`motivated to combine Bian with the Fluhr references because all three
`
`references purportedly disclose similar systems for using query translation to
`
`perform crosslingual information retrieval. Pet. 46. Google further asserts
`
`that it would have been “natural” for an ordinarily skilled artisan to “modify
`
`the system disclosed by Bian to incorporate Fluhr’s query normalization
`
`function,” and that such modification would both maintain the function of
`
`Bian’s system, and yield predictable results. Id.; see also Ex. 1002 ¶ 189.
`
`Google observes that although Bian does not cite to the Fluhr
`
`references themselves, it includes citations to papers presented at the same
`
`20
`
`
`
`IPR2016-00797
`Patent 6,604,101 B1
`
`
`
`symposium as Fluhr ’97, as well as to a dissertation authored by a co-author
`
`of Fluhr ’97. Pet. 46. On this basis, Google asserts that an ordinarily skilled
`
`artisan “would have been motivated to look to the teachings of Fluhr ’97 and
`
`Fluhr ’98 to modify the Bian system, and would reasonably have expected
`
`he or she could do so.” Id. at 46–47; see also Ex. 1002 ¶ 188.
`
`Critically absent from the rationale to combine Bian and the Fluhr
`
`references advanced by Google, however, is any explanation as to why an
`
`ordinarily skilled artisan would look the Fluhr references, which teach an
`
`approach that is “opposite to the one consisting of a translation of the query
`
`followed by a monolingual interrogation” (Ex. 1003, 3), to modify Bian,
`
`which describes a type of system the Fluhr references disparage (Ex. 1006,
`
`4). See DePuy Spine, Inc. v. Medtronic Sofamor Danek, Inc., 567 F.3d
`
`1314, 1326 (Fed. Cir. 2009) (“the ‘predictable result’ discussed in KSR
`
`refers not only to the expectation that prior art elements are capable of being
`
`physically combined, but also that the combination would have worked for
`
`its intended purpose. . . . The opposite conclusion would follow, however, if
`
`the prior art indicated that the invention would not have worked for its
`
`intended purpose or otherwise taught away from the invention.”).
`
`Indeed, Google’s conclusory assertion that Bian and the Fluhr
`
`refere