`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`_______________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`_____________
`
`GOOGLE LLC
`
`Petitioner
`
`v.
`
`VALTRUS INNOVATIONS LIMITED
`
`(record) Patent Owner
`
`
`IPR2022-01545
`Patent No. 6,728,704
`
`
`
`
`PETITION FOR INTER PARTES REVIEW
`UNDER 35 U.S.C. §§ 311-319 AND 37 C.F.R. § 42.200 ET. SEQ
`
`
`
`
`
`
`
`

`

`Patent No. 6,728,704
`
`TABLE OF CONTENTS
`TABLE OF EXHIBITS ............................................................................................. 5
`NOTICE OF LEAD AND BACKUP COUNSEL ..................................................... 6
`NOTICE OF THE REAL-PARTIES-IN-INTEREST ............................................... 6
`NOTICE OF RELATED MATTERS ........................................................................ 6
`NOTICE OF SERVICE INFORMATION ................................................................ 6
`GROUNDS FOR STANDING .................................................................................. 6
`STATEMENT OF PRECISE RELIEF REQUESTED .............................................. 7
`THRESHOLD REQUIREMENT FOR INTER PARTES REVIEW ........................ 7
`I.
`INTRODUCTION ........................................................................................... 7
`A.
`The ’704 Patent Disclosure ................................................................... 7
`CLAIM CONSTRUCTION ............................................................................ 9
`II.
`III. DETAILED EXPLANATION OF THE REASONS FOR
`UNPATENTABILITY .......................................................................... 9
`Ground 1. Claims 1-3, 5-9, 12-14, 17-20, and 23 were Obvious over Bushee
`and Voorhees. ........................................................................................ 9
`A. Overview of the Ground ...................................................................... 10
`1.
`Overview of Bushee .................................................................. 10
`2.
`Overview of Voorhees .............................................................. 12
`Rationale Supporting Obviousness ..................................................... 15
`Graham Factors ................................................................................... 17
`Reasonable Expectation of Success .................................................... 17
`Analogous Art ..................................................................................... 18
`Claim Mapping .................................................................................... 18
`
`B.
`C.
`D.
`E.
`F.
`
`2
`
`

`

`Patent No. 6,728,704
`
`Ground 2. Claims 3-4, 9-10, 14-15, and 20-21 were Obvious over Bushee,
`Voorhees and Koppel .......................................................................... 37
`A. Overview of the Ground ...................................................................... 37
`B.
`Overview of Koppel ............................................................................ 38
`C.
`Graham Factors ................................................................................... 39
`D.
`Rationale (Motivation) for the Combination ....................................... 39
`E.
`Reasonable Expectation of Success .................................................... 39
`F.
`Analogous Art ..................................................................................... 40
`G.
`Claim Mapping .................................................................................... 40
`Ground 3. Claims 1-3, 5-9, 12-14, 17-20, and 23 are Obvious over Voorhees,
`Bushee and Tso. .................................................................................. 41
`Effective Prior Art Dates ..................................................................... 41
`Overview of the Ground ...................................................................... 42
`1.
`Overview of Tso ........................................................................ 42
`Graham Factors ................................................................................... 44
`C.
`D. Analogous Art ..................................................................................... 44
`E.
`Rationale for the Combination ............................................................ 45
`F.
`Reasonable Expectation of Success .................................................... 45
`G.
`Claim Mapping .................................................................................... 45
`Ground 4. Claims 3-4, 9-10, 14-15, and 20-21 were Obvious over Voorhees in
`view of Bushee, Tso, and Koppel ....................................................... 66
`Explanation of the Ground .................................................................. 66
`H.
`INSTITUTION UNDER 35 U.S.C. § 325(d) ................................................ 67
`INSTITUTION UNDER 35 U.S.C. § 314 .................................................... 69
`
`IV.
`V.
`
`A.
`B.
`
`3
`
`

`

`Patent No. 6,728,704
`
`VI. CONCLUSION .............................................................................................. 71
`CERTIFICATE OF SERVICE ................................................................................ 73
`CERTIFICATE OF WORD COUNT ...................................................................... 74
`
`
`
`4
`
`

`

`Patent No. 6,728,704
`
`TABLE OF EXHIBITS
`
`
`Description
`U.S. Patent No. 6,728,704 (“the ’704 patent”).
`Declaration of Prof. Padhraic Smyth.
`C.V. of Prof. Padhraic Smyth.
`U.S. Pat. No. 6,711,569 (“Bushee”).
`U.S. Pat. No. 5,864,846 (“Voorhees”).
`U.S. Pat. No. 6,385,602 (“Tso”).
`File History of U.S. Pat. App. Ser. No. 09/940,600 (“the ’704
`patent file history”).
`Voorhees, et al., “Multiple Search Engines in
`Database Merging,” In Proceedings of the Second ACM
`International Conference on Digital Libraries (DL'97),
`1997.
`Voorhees, et al., “The Collection Fusion Problem,”
`In Proceedings of the Third Text Retrieval Conference
`(TREC-3), 1995.
`United States District Courts — National Judicial Caseload
`Profile,
`March
`31,
`2022,
`available
`at
`https://www.uscourts.gov/statistics/table/na/federal-court-
`management-statistics/2022/03/31-1.
`Docket printout for Valtrus Innovations, Ltd. v. Google LLC, No.
`3:22-cv-00066-N (N.D. Tex.).
`Dkt. 51, Scheduling Order, Valtrus Innovations, Ltd. v. Google
`LLC, No. 3:22-cv-00066-N (N.D. Tex. Apr. 27, 2022).
`U.S. Pat. No. 7,257,766 (“Koppel”).
`
`
`Exhibit No.
`1001
`1002
`1003
`1004
`1005
`1006
`1007
`
`1008
`
`1009
`
`1010
`
`1011
`
`1012
`
`1013
`
`
`
`5
`
`

`

`Patent No. 6,728,704
`
`Petitioner respectfully requests inter partes review under 35 U.S.C. §311 of
`
`claims 1-23 of U.S. Pat. No. 6,728,704 (“the ’704 patent”).
`
`NOTICE OF LEAD AND BACKUP COUNSEL
`Lead Counsel
`Backup Counsel
`Matthew A. Smith
`Andrew S. Baluch
`Reg. No. 49,003
`Reg. No. 57,503
`SMITH BALUCH LLP
`SMITH BALUCH LLP
`700 Pennsylvania Ave. SE, Ste 2060
`700 Pennsylvania Ave. SE, Ste 2060
`Washington, DC 20003
`Washington, DC 20003
`(202) 669-6207
`(202) 880-2397
`smith@smithbaluch.com
`baluch@smithbaluch.com
`
`
`NOTICE OF THE REAL-PARTIES-IN-INTEREST
`The real-party-in-interest for this petition is Google LLC (“Google”).
`
`NOTICE OF RELATED MATTERS
`The ’704 patent is asserted in the following active litigation: Valtrus
`
`
`
`Innovations Ltd v. Google LLC, Case No. 3-22-cv-00066 (N.D. Tex.).
`
`NOTICE OF SERVICE INFORMATION
`Please address all correspondence to the lead counsel at the addresses shown
`
`above.
`
`Petitioners
`
`consent
`
`to
`
`electronic
`
`service
`
`by
`
`email
`
`at:
`
`smith@smithbaluch.com, baluch@smithbaluch.com.
`
`GROUNDS FOR STANDING
`Petitioner hereby certifies that the patent for which review is sought is
`
`available for inter partes review, and that the Petitioner is not barred or estopped
`
`from requesting a inter partes review on the grounds identified in the petition.
`
`6
`
`

`

`Patent No. 6,728,704
`
`STATEMENT OF PRECISE RELIEF REQUESTED
`Petitioner respectfully requests that claims 1-23 of the ’704 patent be canceled
`
`based on the following grounds:
`
`Ground 1: Claims 1-3, 5-9, 11-14, 16-20, and 22-23 were Obvious over
`
`Bushee and Voorhees.
`
`Ground 2: Claims 3-4, 9-10, 14-15, and 20-21 were Obvious over Bushee,
`
`Voorhees and Koppel.
`
`Ground 3: Claims 1-3, 5-9, 12-14, 17-20, and 23 were Obvious over
`
`Voorhees, Bushee and Tso.
`
`Ground 4: Claims 3-4, 9-10, 14-15, and 20-21 were Obvious over Voorhees,
`
`Bushee, Tso, and Koppel.
`
`THRESHOLD REQUIREMENT FOR INTER PARTES REVIEW
`As shown in the Grounds set forth below, the information presented in the
`
`instant petition, if unrebutted, demonstrates that “it is more likely than not that at
`
`least 1 of the claims challenged in the petition is unpatentable.” 35 U.S.C. § 314(a).
`
`I.
`
`INTRODUCTION
`A. The ’704 Patent Disclosure
`The ’704 patent relates to a particular method for arranging Web search
`
`results. (Ex. 1001, Abstract)(Ex. 1002, ¶30). The search results arise from a query
`
`provided to multiple search engines. (Ex. 1001, 1:37-52)(Ex. 1002, ¶30). Each
`
`search engine returns a result list that includes links to Web pages. (Id.). Each Web
`
`7
`
`

`

`Patent No. 6,728,704
`
`page has a relevance score, which is computed using known methods. (Ex. 1001,
`
`5:56-65)(Ex. 1002, ¶30).
`
`When the ’704 patent system receives lists of results from different search
`
`engines, it seeks to present them as a single list to a user. (Ex. 1001, 6:6-28)(Ex.
`
`1002, ¶31). To do so, the system takes a subset of each list, and then assigns each
`
`search engine a score (a “representative value”) to represent the overall relevance of
`
`the search engine’s results. (Id.). One example of such a representative value is the
`
`average of the scores of each entry in the subset. (Id.).
`
`The representative value is used to sort the results from multiple search
`
`engines into a single list. (Ex. 1002, ¶32). To do this, the ’704 patent teaches two
`
`methods. (Id.). As explained in the ’704 patent:
`
`“In the first embodiment, entries are merged by selecting the list with
`the highest representative value (e.g., the highest average scoring
`value). The first entry on the list that has not already been selected is
`then picked. That list’s representative value is then decremented by
`a fixed amount and the process is repeated until all entries have been
`picked….In the second embodiment…[e]ach list is assigned a
`probability value equal to its representative value’s percentage of the
`total representative values for all lists. Lists are then selected
`according to their probability value, with lists having higher
`probability values being more likely to be selected.”
`
`(Ex. 1001, 6:8-24)(Ex. 1002, ¶32).
`
`8
`
`

`

`Patent No. 6,728,704
`
`II. CLAIM CONSTRUCTION
`Petitioner does not believe that the Petition presents an issue of claim
`
`construction for the prior art presented in the Grounds.
`
`III. DETAILED EXPLANATION OF THE REASONS FOR
`UNPATENTABILITY
`
`Ground 1. Claims 1-3, 5-9, 12-14, 17-20, and 23 were Obvious over Bushee
`and Voorhees.
`
`Claims 1-3, 5-9, 12-14, 17-20, and 23 are obvious under pre-AIA 35 U.S.C.
`
`§ 103(a) over U.S. Pat. No. 6,711,569 (“Bushee”)(Ex. 1004) in view of U.S. Pat.
`
`5,864,846 (“Voorhees”)(Ex. 1005).
`
`Neither Bushee nor Voorhees was of record during the prosecution of any
`
`application leading to the ’704 patent.
`
`Bushee is a U.S. patent that issued on March 23, 2004 from an application
`
`filed in the United States on July 24, 2001. It is thus prior art under pre-AIA 35
`
`U.S.C. §102(e).
`
`Voorhees is a U.S. patent that issued on January 26, 1999, and is thus prior art
`
`under pre-AIA 35 U.S.C. §§102(b).
`
`9
`
`

`

`Patent No. 6,728,704
`
`A. Overview of the Ground
`Bushee teaches all of claim 1 except (under the proper claim interpretation1)
`
`merging the results “in a predetermined manner based on the representative
`
`value...wherein
`
`the representative value varies
`
`in accordance with [the]
`
`predetermined manner” (claim limitation [1h]). However, Voorhees teaches a
`
`method of sorting pages into a merged result list that uses a representative value,
`
`wherein the representative value varies in accordance with the predetermined
`
`manner. (Ex. 1002, ¶¶40-41).
`
`This ground posits that it would have been obvious to use Voorhees’s method
`
`of sorting pages into a merged list, within Bushee’s multi-search-engine system. (Ex.
`
`1002, ¶45). This ground also makes a number of minor obviousness assertions,
`
`including, for example, that it was obvious to use a “scoring function” to determine
`
`a numerical score for a document, and that it would have been obvious, each time a
`
`page is assigned from an individual search engine’s results to the final result list, to
`
`decrement the number of pages left to be assigned from that search engine’s results.
`
`1. Overview of Bushee
`Bushee teaches a “method for automatic selection of databases for improving
`
`
`1 Bushee anticipates claim 1 under Valtrus’ apparent litigation claim construction
`
`positions.
`
`10
`
`

`

`Patent No. 6,728,704
`
`the efficiency of data capture and management systems.” (Ex. 1004, Abstract)(Ex.
`
`1002, ¶44). While Bushee often speaks in terms of “databases” having “documents”,
`
`it is clear that Bushee’s databases are actually search engines operating on the World
`
`Wide Web. (Ex. 1004, 2:1-7)(“the terms search engine and database are also used
`
`synonymously”)(Ex. 1002, ¶44).
`
`Bushee explains the method of the invention in relation to Fig. 2, which is
`
`reproduced here:
`
`
`
`11
`
`

`

`Patent No. 6,728,704
`
`(Ex. 1004, Fig. 2)(Ex. 1002, ¶44). As shown above, Bushee obtains a query from a
`
`user, selects databases (search engines) to send the query to, transmits the query,
`
`receives results from the search engines, and selects a subset of the results (the “first
`
`N results”). (Ex. 1004, 3:64-5:5)(Ex. 1002, ¶44). Then, Bushee scores each result
`
`in each subset, takes the average of the scores for each search engine, and uses the
`
`average score of each database to rank the databases for presentation. (Ex. 1004,
`
`4:47-5:33)(Ex. 1002, ¶44).
`
`Bushee explains that—just like in the ’704 patent—the average score
`
`represents a measure of the relevancy of each database (search engine) to the user’s
`
`query:
`
`“The numerical scores for each document retrieved from a database
`are averaged together, and this averaged score is then assigned to
`the database as an indication of relevance of that database to the
`user's query.”
`
`(Ex. 1004, 5:18-28)(Emphasis added)(Ex. 1002, ¶45).
`
`2. Overview of Voorhees
`Voorhees—like Bushee and the ’704 patent—teaches a system for receiving
`
`search results from multiple search engines. (Ex. 1005, Abstract)(Ex. 1002, ¶46).
`
`There are two aspects to the Voorhees technology that are relevant to the
`
`present Petition: a method of assigning a relevance weight to each search engine
`
`that returns results, and a method of combining (or “fusing”) the results of multiple
`
`12
`
`

`

`Patent No. 6,728,704
`
`search engines into a single search result for the user. The former method (assigning
`
`a weight) is most relevant to Ground 2, while the latter (fusing the results into a
`
`single document) is relevant to both Ground 1 and Ground 2.
`
`Voorhees teaches fusing search results from different search engines into a
`
`single search results page by estimating the relevance of each search engine’s results.
`
`To do this, Voorhees proposes two alternate methods, called (1) the average relevant
`
`document distribution method and (2) the centroid method. The present Petition
`
`(especially in Ground 2) focuses on the centroid method. (Ex. 1005, 4:30-6:29)(Ex.
`
`1002, ¶48). In the centroid method, each search engine has already performed a
`
`number of training queries with known results. (Ex. 1005, 2:43-51)(Ex. 1002, ¶48).
`
`Both the training queries and the results they return can differ from search engine to
`
`search engine. (Ex. 1002, ¶48). At each search engine, the training queries are
`
`divided into clusters, that is, groups of queries that have similar meanings. (Ex.
`
`1004, 4:41-63, 5:35-56)(Ex. 1002, ¶48).
`
`These clusters are used to process a user query. When a user query is received
`
`by a search engine, the search engine compares the query with each cluster’s
`
`centroid. (Ex. 1005, 5:65-67)(Ex. 1002, ¶¶49-50). A cluster centroid is basically an
`
`average of all queries that make up the cluster. (Ex. 1005, 5:45-47)(Ex. 1002, ¶¶49-
`
`50). The cluster centroid that is most similar to the user’s query is selected at each
`
`search engine, and a search-engine weight (ws) is obtained. (Ex. 1005, 5:65-6:9)(Ex.
`
`13
`
`

`

`Patent No. 6,728,704
`
`1002, ¶¶49-50). The search-engine weight ws is the mean number of relevant pages
`
`above a certain rank expected to be returned by a query similar to the cluster. (Ex.
`
`1005, 5:57-64)(Ex. 1002, ¶49). Because the search-engine weight ws can be
`
`different for each search engine, the weight ‘w’ is denoted by a subscript ‘s’
`
`indicating the search-engine number. (Ex. 1005, 5:2-3)(Ex. 1002, ¶¶49-50).
`
`Once the weight for each search engine (ws) is returned, Voorhees assigns a
`
`number of pages (λs) to each search engine (where the subscript s again denotes the
`
`particular search engine). (Ex. 1005, 4:67-5:11, 6:10-18)(Ex. 1002, ¶51). To do
`
`this, Voorhees takes a total number of pages N to be retrieved, and divides up the
`
`pages proportionally to the search engine weights. (Ex. 1005, 4:67-5:11)(Ex. 1002,
`
`¶51).
`
`Once Voorhees has selected the number of pages to take from each search
`
`engine, Voorhees fuses the result sets into a combined-results page using a semi-
`
`random process, which Voorhees calls rolling a “biased c-faced die”. (Ex. 1005,
`
`6:19-29, 4:1-29)(Ex. 1002, ¶52). This process selects a page using a random number
`
`that is weighted by the number of pages remaining to be selected from each search
`
`engine. (Id.). Thus, when a page is selected from one search engine’s results, the
`
`number of pages remaining to be selected from that search engine is decremented,
`
`and the likelihood that the search engine will be selected to provide the next page
`
`for the combined results will decrease. (Ex. 1005, 4:19-29)(Ex. 1002, ¶52).
`
`14
`
`

`

`Patent No. 6,728,704
`
`B. Rationale Supporting Obviousness
`While Bushee would anticipate claim 1 under Valtrus’ litigation claim
`
`constructions, under a proper construction of the claims, Bushee would not teach the
`
`final limitation (limitation [1h]) of claim 1, which reads: “[1h] wherein the
`
`representative value varies in accordance with predetermined manner.” However,
`
`Voorhees teaches this limitation, because its representative value (the number of
`
`pages, λs) changes each time a web page is allocated from one search engine’s results
`
`to the combined results.
`
` It would have been obvious to use, in the system of Bushee, Voorhees’
`
`method for fusing search results from multiple search engines into a single page.
`
`(Ex. 1002, ¶54). Specifically, it would have been obvious to use Bushee’s average
`
`scores as search-engine weights ws (Ex. 1005, 4:67-5:3) in Voorhees’ fusion method,
`
`allowing Voorhees to calculate the number of pages λs allocated to each search
`
`engine using the weights provided by Bushee’s system, and to use Voorhees’ semi-
`
`random method for combining results from multiple search engines into a single list.
`
`(Ex. 1002, ¶54).
`
`The combination would have been obviously advantageous. Bushee teaches
`
`a rudimentary method for fusion, namely, placing the results of each search engine
`
`into a page together as a block, with each block in the rank order of the search
`
`engines. (Ex. 1004, 5:29-33)(Ex. 1002, ¶55). Because the results of each search
`
`15
`
`

`

`Patent No. 6,728,704
`
`engine appear together with other results of that search engine, the best results from
`
`the second-ranked search engine are placed in the combined list after the worst
`
`results of the first-ranked search engine. (Ex. 1002, ¶55). As a result, a POSITA
`
`would have understood that a user is unlikely to be presented with the best search
`
`overall results in ranked order (regardless of source), as would ordinarily be
`
`expected from a Web search. (Ex. 1002, ¶55).
`
`The fusion method of Voorhees would have been seen as an improvement
`
`over that of Bushee because it provides for a true fusion of results from each search
`
`engine into a combined list, where the top results of each search engine are likely to
`
`appear near the top of the list. (Ex. 1002, ¶56). This would have allowed the user
`
`to view a single ranking of results, with the best results from each search engine near
`
`the top of the ranking, thus achieving the benefit of using multiple search engines in
`
`a manner transparent to the user. (Ex. 1005, 1:52-2:2)(Ex. 1002, ¶56). This
`
`advantage would have been apparent and would have motivated a POSITA to make
`
`the combination.
`
`Furthermore, the combination was obvious under KSR Int’l Co. v. Teleflex,
`
`Inc., 550 U.S. 398, 416-21 (2007). Bushee represented a known prior art disclosure
`
`for searching the Web while harnessing the power of multiple search engines. While
`
`Bushee teaches a rudimentary method for fusing search results from multiple search
`
`engines, Voorhees teaches a known and more sophisticated alternative for doing so,
`
`16
`
`

`

`Patent No. 6,728,704
`
`which could have been implemented using relatively simple functions on a digital
`
`computer. (Ex. 1002, ¶57). The combination would thus have been within the skill
`
`of a POSITA, and would have led to predictable results (Ex. 1002, ¶57), and would
`
`have been obvious under KSR. See KSR 550 U.S. at 417 (2007)(“a court must ask
`
`whether the improvement is more than the predictable use of prior art elements
`
`according to their established functions.”).
`
`More specific detail regarding rationale supporting obviousness will be
`
`provided in the claim mapping section below, under each element where relevant.
`
`C. Graham Factors
`The level of ordinary skill encompassed a person having a Bachelor’s Degree
`
`in Computer Science or a related field and five years of experience in search
`
`technology, where a higher level of education may substitute for experience and vice
`
`versa. (Ex. 1002, ¶¶26-28).
`
`The scope and content of the prior art are discussed throughout the Ground.
`
`The differences between the prior art and the claims are discussed in the
`
`“Overview of the Combination” and in the claim mapping, below.
`
`Petitioner is not aware of any secondary considerations that would make an
`
`inference of non-obviousness more likely.
`
`D. Reasonable Expectation of Success
`A person of ordinary skill in the art in the relevant timeframe would have had
`
`17
`
`

`

`Patent No. 6,728,704
`
`a reasonable expectation of success in using the prior art in the manner discussed in
`
`this petition. (Ex. 1002, ¶60). As Dr. Smyth explains, the art was relatively
`
`predictable in the relevant timeframe (August 2001). (Ex. 1002, ¶60). A POSITA
`
`would have been able to make any necessary modifications to implement the
`
`Ground, and in particular would have been able to use the average scores of Bushee
`
`as weights in the process of Voorhees, while using the fusion method of Voorhees
`
`to integrate results from multiple search engines in Bushee. (Ex. 1002, ¶60).
`
`E. Analogous Art
`Bushee and Vorhees are analogous art because they are in the same field as
`
`the ’704 patent (information retrieval and search engine technology). (Ex. 1001,
`
`1:7-8)(Ex. 1005, 1:6-7)(Ex. 1004, 1:41)(Ex. 1002, ¶61). Furthermore, the methods
`
`of Voorhees and Bushee would have been reasonably pertinent to the problems
`
`facing the named inventors, for example, the problem of merging results from
`
`multiple search engines. (Ex. 1001, 1:9-10)(Ex. 1005, 1:10-11)(Ex. 1004, 5:29-
`
`33)(Ex. 1002, ¶61). See Wyers v. Master Lock Co., 616 F.3d 1231, 1238 (Fed. Cir.
`
`2010)(“The Supreme Court’s decision in KSR [cite omitted], directs us to construe
`
`the scope of analogous art broadly….”).
`
`F. Claim Mapping
`This section maps the challenged claims to the relevant disclosures of
`
`Voorhees and Bushee, where the claim text appears in bold-italics, and the relevant
`
`18
`
`

`

`Patent No. 6,728,704
`
`mapping follows the claim text. The Petitioner has added numbering and lettering
`
`in brackets (e.g. 1[a], [1b]) to certain claim elements, to facilitate the discussion.
`
`CLAIM 1
`
`“1[a]. A method of merging result lists from multiple search
`engines, said method comprising:”
`
`To the extent the preamble is limiting, the combination of Bushee and
`
`Voorhees teaches a method of merging result lists from multiple search engines, as
`
`shown in the discussion under elements [1b] - [1h] below. (Ex. 1002, ¶63).
`
`“[1b] transmitting a query to a set of search engines;”
`
`Bushee teaches transmitting a query to a set of search engines. Bushee
`
`states:
`
`“The query input means 30 of the system 10 is used for receiving a
`plurality of queries from a user and transferring the plurality of
`queries to a plurality of databases 4.”
`
`(Ex. 1004, 3:65-67)(Emphasis added)(Ex. 1002, ¶64). As previously noted,
`
`Bushee’s references to “databases” are actually references to search engines. (Ex.
`
`1004, 2:5-7)(Ex. 1002, ¶65).
`
`“[1c] receiving in response to said query a result list from each
`search engine of said set of search engines, each result list including
`one or more entries;”
`
`As discussed under limitation [1b], above, Bushee teaches receiving a
`
`19
`
`

`

`Patent No. 6,728,704
`
`document in response to a query from each search engine, because it teaches
`
`“capturing, storing and scoring a plurality of responsive documents 70 (such as,
`
`for example, web pages) returned by each one of the plurality of databases 4 in
`
`response to the user’s query.” (Ex. 1004, 4:1-5)(Emphasis added)(Ex. 1002, ¶66).
`
`The responsive web pages are result lists including one or more entries,
`
`because they “provide URLs for responsive documents.” (Ex. 1004, 4:53-57)(Ex.
`
`1002, ¶67). That is, the searches produce web pages that have links (URLs) to
`
`responsive documents, which are the one or more entries. (Id.).
`
`“[1d] selecting a subset of entries from each result list to form a set
`of selected entries;”
`
`Bushee teaches selecting a subset of entries from each result list to form a
`
`set of selected entries, because it teaches, in connection with Fig. 2, “pull[ing] [the]
`
`First N Results from Each Database”. (Ex. 1004, Fig. 2)(Ex. 1002, ¶68). The
`
`relevant portion of Fig. 2 is shown below, with highlighting added:
`
`20
`
`

`

`Patent No. 6,728,704
`
`
`(Ex. 1004, Fig. 2)(Ex. 1002, ¶68). Bushee explains that “[a]n evaluation parameter
`
`may be used to define a maximum number of responsive documents to be captured
`
`from each one of the plurality of databases.” (Ex. 1004, 4:61-67)(Ex. 1002, ¶68).
`
`Thus, Bushee teaches selecting a subset of entries from each result list. (Id.). The
`
`combination of these subsets form a set of selected entries. (Ex. 1002, ¶68).
`
`21
`
`

`

`Patent No. 6,728,704
`
`“[1e] assigning to each selected entry of said set of selected entries a
`scoring value according to a scoring function;”
`
`Bushee next renders obvious assigning to each selected entry of said set of
`
`selected entries a scoring value according to a scoring function, as shown in Fig.
`
`2, the relevant portion of which is shown here:
`
`(Ex. 1004, Fig. 2)(Ex. 1002, ¶69). Bushee explains:
`
`
`
`“The evaluation portion 40 determines a page score for a
`representation of each one of the responsive documents 70
`associated with each one of the plurality of databases 4.”
`
`(Ex. 1004, 4:10-13)(Emphasis added)(Ex. 1002, ¶69). It would have been obvious
`
`to a POSITA that the evaluation portion 40 used a mathematical method, embodied
`
`in software, to “determine a page score”, and thus used a scoring function. This is
`
`implied by the word “score”, and was a well-known technique in the relevant
`
`timeframe. (Ex. 1002, ¶69)(Ex. 1005, 1:13-17).
`
`22
`
`

`

`Patent No. 6,728,704
`
`“[1f] assigning to each subset a representative value according to
`the scoring values assigned to its entries; and”
`
`Bushee in view of Voorhees renders this element obvious. Specifically, it
`
`would have been obvious to calculate a number of pages to select from each search
`
`engine (a representative value), per Voorhees, based on the search engine
`
`weighting provided by Bushee. (Ex. 1002, ¶70).
`
`First, Bushee teaches calculating a search-engine weight that is the average of
`
`scores of each individual search result provided by that search engine. (Ex. 1004,
`
`4:15-20)(Ex. 1002, ¶71). Bushee states:
`
`“The evaluation portion further determines an averaged score for
`each one of the plurality of databases 4 based upon an average of
`each one of the page scores. The averaged score is used to evaluate
`the relevancy of the database 4 to the user’s query.”
`
`(Ex. 1004, 4:15-20)(Ex. 1002, ¶71).
`
`From there, it would have been obvious, per Voorhees, to calculate a number
`
`of pages to select from each search engine. (Ex. 1002, ¶72). Specifically, Voorhees
`
`teaches (in connection with its “centroid” method) to obtain a weight for each search
`
`engine that returns a result, and using that weight, to calculate a number of pages to
`
`return from each search engine. Voorhees states:
`
`“The set of weights returned over all the search engines is used to
`apportion the final retrieved set such that when N pages are to be
`returned and ws is the weight returned by engine s, (ws/Σws)*N
`
`23
`
`

`

`Patent No. 6,728,704
`
`(rounded appropriately) documents are selected from engine s. For
`example, assume the total number of pages to be retrieved is 100,
`and there are 5 five search engines. If the weights returned by the
`engines are 4, 3, 3, 0, 2, then the first 33 pages returned by engine 1
`would be selected, the first 25 pages from each of engines 2 and 3
`would be selected, no pages would be selected from engine 4, and
`the 10 first 17 pages from engine 5 would be selected. However, if
`the weights returned were 4, 8, 4, 0, 0 then 25 pages would be
`selected from each of engines 1 and 3, and 50 pages would be
`selected from engine 2.”
`
`(Ex. 1005, 4:67-5:13)(Ex. 1002, ¶72).
`
`In the combination, the number of pages to retrieve from each search engine
`
`(denoted λs in Voorhees) is the representative value that is assigned to each
`
`subset. (Ex. 1002, ¶73). It is a representative value because the number of pages
`
`represents the probable relevance of the search engine, because it is proportional to
`
`the search engine weight. (Ex. 1005, 4:64-5:15)(Ex. 1002, ¶73). Furthermore, the
`
`representative value is assigned according to the scoring values assigned to its
`
`entries because, in the combination, it is based on the average of the scoring values
`
`assigned to each of its entries, as shown in this diagram, created by the Petitioner:
`
`
`
`24
`
`

`

`Patent No. 6,728,704
`
`(Ex. 1002, ¶73).
`
`It would have been obvious to calculate a number of pages to retrieve from
`
`each search engine in Bushee, by using the method of Voorhees (selecting the
`
`number of pages to retrieve proportional to the weight (i.e. the average score for each
`
`search engine), for the reasons provided above in the section entitled “Rationale
`
`Supporting Obviousness”, beginning on page 7, above. (Ex. 1002, ¶74).
`
`“[1g] producing a merged list of entries in a predetermined manner
`based on the representative value assigned to each result list,”
`
`Voorhees teaches producing a merged list of entries in a predetermined
`
`manner based on the representative value assigned to each result list.
`
`First, Voorhees generally teaches a method to produce a merged list of
`
`entries from results obtained from separate search engines. (Ex. 1005, Abstract)(Ex.
`
`1002, ¶76).
`
`Voorhees performs the merger in a predetermined manner based on the
`
`representative value assigned to each result list. Specifically, Voorhees teaches:
`
`“4. Rank[ing] pages in the retrieved set probabilistically using a
`biased c-faced die. (a) To select the document to be in the next rank
`r of the final ranking, roll a c-faced die that is biased by the number
`of pages remaining to be placed in the final ranking from each of the
`engines. Select the engine whose number corresponds to the die roll
`and place the next page from that engine's ranking into the final
`ranking. (b) Repeat until all N pages have been placed in the final
`
`25
`
`

`

`Patent No. 6,728,704
`
`ranking.”
`
`(Ex. 1005, 6:19-29)(Ex. 1002, ¶77). The “final rankin