`
`Filed on Behalf of NETWORK-1 TECHNOLOGIES, INC.
`
`By: Charles R. Macedo (Reg. No. 32,781)
`Brian A. Comack (Reg. No. 45,343)
`Amster, Rothstein & Ebenstein LLP
`90 Park Avenue
`New York, NY 10016
`Telephone: (212) 336–8074
`Facsimile: (212) 336–8001
`cmacedo@arelaw.com
`N1-Google-IPR@arelaw.com
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`____________________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________________
`
`GOOGLE INC.,
`Petitioner,
`
`v.
`
`NETWORK-1 TECHNOLOGIES, INC.,
`Patent Owner.
`
`_______________
`
`Case IPR2015-00345
`Patent 8,205,237
`____________________
`
`
`
`
`
`
`PATENT OWNER’S PRELIMINARY RESPONSE TO THE PETITION
`FOR INTER PARTES REVIEW OF U.S. PATENT NO. 8,205,237 UNDER 35
`U.S.C. § 313 PURSUANT TO 37 C.F.R. § 42.107
`
`
`
`
`600885.2
`
`
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`Table of Contents
`I.
`Introduction. ..................................................................................................... 1
`Overview of the ‘237 Patent. ........................................................................... 3
`II.
`III. Claim Constructions. ....................................................................................... 5
`A.
` “non-exhaustive search.” ...................................................................... 5
`B.
`“sublinear.” ............................................................................................ 9
`C.
` “neighbor” / “near neighbor.” ............................................................ 10
`D.
`“nearest neighbor.” .............................................................................. 11
`E.
`“approximate nearest neighbor search” .............................................. 12
`IV. Ground 1: Anticipated by Iwamura (Ex. 1012) fails. ................................... 13
`A. Overview of Iwamura. ......................................................................... 13
`B.
`Ground 1 fails because Iwamura does not disclose the claimed
`“non-exhaustive,” “sublinear,” and “approximate nearest”
`search limitations. ................................................................................ 14
`1.
`Iwamura does not disclose the claimed non-exhaustive,
`sublinear, or approximate nearest elements but rather
`teaches an exhaustive, linear search. ........................................ 14
`Response to Petitioner’s arguments. ......................................... 16
`2.
`V. Ground 4: Anticipated by Ghias (Ex. 1010) fails. ........................................ 20
`A. Overview of Ghias. .............................................................................. 20
`B.
`Ground 4 fails because Ghias does not disclose the claimed
`sublinear limitations found in each challenged independent
`claim. ................................................................................................... 21
`VI. Ground 5: Anticipated by Wood (Ex.1015) fails. ........................................ 25
`A. Overview of Wood. ............................................................................. 25
`B.
`Ground 5 fails because Wood does not disclose the claimed
`“approximate nearest neighbor search.” ............................................. 26
`VII. Ground 2: Obvious over Levy (Ex. 1013) in view of Arya (Ex. 1006)
`fails. ................................................................................................................ 29
`
`
`
`600885.2
`
`
`
`i
`
`
`
`
`
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`A. Overview of Levy. ............................................................................... 29
`B. Overview of Arya. ............................................................................... 32
`C. Ground 2 fails because the references fail to teach claim
`elements and teach away from the proposed combination. ................ 34
`1.
`Levy and Arya teach away from combining these
`references. ................................................................................. 35
`2. The two asserted motivations to combine Levy and Arya
`presented in the Petition both fail. ............................................ 40
`VIII. Ground 3: Obvious Iggulden (Ex. 1011) in view of Böhm (Ex. 1007)
`fails. ................................................................................................................ 43
`A. Overview of Iggulden. ......................................................................... 43
`B.
`Overview of Böhm. ............................................................................. 45
`C.
`Ground 3 fails because Iggulden and Böhm teach away from
`the proposed combination. .................................................................. 46
`1.
`Iggulden and Böhm teach away from combining the
`references. ................................................................................. 47
`Petitioner’s asserted motivations to combine Iggulden
`and Böhm presented in the Petition fail. ................................... 51
`IX. Ground 6: Obvious over Iwamura (Ex. 1012) in view of Chen (Ex.
`1008) fails. ..................................................................................................... 53
`X. Ground 7: Obvious over Levy (Ex. 1013) in view of Arya (Ex. 1006)
`and Chen (Ex. 1008) fails. ............................................................................. 54
`XI. Conclusion. .................................................................................................... 55
`
`
`
`
`2.
`
`
`
`
`
`600885.2
`
`
`
`ii
`
`
`
`
`
`
`
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`TABLE OF AUTHORITIES
`
` Page(s)
`
`Cases
`CallCopy, Inc. v. Verint Americas, Inc.,
`IPR2013-00492, Paper 14 (PTAB Feb. 5, 2014) .................................................... 54
`Glaxo Group Ltd. v. Apotex, Inc.,
`376 F.3d 1339 (Fed. Cir. 2004) ............................................................................... 31
`In re Gordon,
`733 F.2d 900 (Fed. Cir. 1984) ................................................................................. 36
`International Securities Exchange v. Chicago Board Options Exchange,
`IPR2014-00097, Paper 12 (PTAB May 22, 2014) ........................................... passim
`McGinley v. Franklin Sports, Inc.,
`262 F.3d 1339 (Fed. Cir. 2001) ......................................................................... 36, 48
`Nystrom v. Trex Co.,
`424 F.3d 1136 (Fed. Cir. 2005) ............................................................................... 11
`Smart Modular Technologies v. Netlist, Inc.,
`IPR2014-01371, Paper 12 (PTAB March 13, 2015) ................................................. 8
`Statutes
`35 U.S.C. § 312 ............................................................................................................. 12
`35 U.S.C. § 316(e)......................................................................................................... 35
`Regulations
`37 C.F.R. § 42.104(b)(3) ............................................................................................... 12
`37 C.F.R. § 42.104(b)(4) ................................................................................................. 3
`
`
`
`
`
`
`600885.2
`
`
`
`
`
`iii
`
`
`
`
`
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`
`
`Exhibit List
`
`EXHIBIT NO.
`Exhibit 2001
`
`Exhibit 2002
`
`Exhibit 2003
`
`
`
`DESCRIPTION
`“Brute-force search”—http://en.wikipedia.org/wiki/Brute-
`force_search (3/19/2015)
`
`U.S. Patent No. 8,447,762 (Brendel)
`
`U.S. Patent No. 7,167,984 (Graveman)
`
`
`
`
`600885.2
`
`
`
`iv
`
`
`
`
`
`
`I.
`
`Introduction.
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`The references asserted in the Petition fall into two general categories:
`
`Category 1 consists of patent references—Ghias (Ex. 1010), Wood (Ex.
`
`1015), Levy (Ex. 1013), Iggulden (Ex. 1011), Iwamura (Ex. 1012), and Chen (Ex.
`
`1008). Rather than teaching to identify a work using the types of searches claimed
`
`in the ‘237 Patent (addressed below), these references explicitly teach using a
`
`different type of search to identify matches:
`
` Ghias teaches an exhaustive search that “compare[s] all the songs” in the
`
`database (Ghias, 5:66-6:2);
`
` Wood teaches an exhaustive search that compare a “specimen” with “each entry
`
`in the pattern library” (Wood, 6:8-23);
`
` Iwamura teaches an exhaustive search that compares the melody to be identified
`
`with each record such that “[t]he reference melody that gives the least different
`
`is returned as a search result” (Iwamura, 7:53-55);
`
` Levy teaches using exact match comparisons rather than neighbor searches to
`
`identify a match using a extracted hashed fingerprint (Levy, 9:42-61);
`
` Iggulden teaches using exact match comparisons rather than neighbor searches
`
`to identify a match using an extracted hash as a “unique signature” (Iggulden,
`
`9:55-57, 9:59-62); and
`
`600885.2
`
`1
`
`
`
` Chen provides no details as to how any search is performed.
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`Category 2 consists of academic articles—Arya (Ex. 1006), and Böhm (Ex.
`
`1007)—that identify neighbor searches, but do not disclose other claim elements.
`
`The Petition presents seven overlapping and cumulative grounds. Three
`
`grounds are based on single references (grounds 1, 4, and 5); four grounds are
`
`based on combinations (grounds 2, 3, 6 and 7).
`
`For the single reference grounds (grounds 1, 4, and 5), the Petition fails to
`
`satisfy its burden of demonstrating that claimed elements are found in the asserted
`
`references. Moreover, while Petitioner identifies constructions to be applied, it
`
`fails to map key constructions onto the asserted art. For example, Petitioner
`
`construes “identify a neighbor” as “identify a close, but not necessarily exact
`
`match.” Pet. 6. But after asserting its construction, Petitioner’s construction
`
`appears nowhere in its Petition.
`
`For the combination grounds (2, 3, 6, and 7), not only are elements missing
`
`from the proposed combinations, the alleged combinations, based on flawed
`
`premises, would not work and the references teach away from such combinations
`
`rather than provide any motivation to combine. The proposed combinations of the
`
`academic articles (Arya and Böhm) with the patent references (Levy and Iggulden)
`
`rely on the hashing feature extraction methods taught in Levy and Iggulden. Arya
`
`600885.2
`
`2
`
`
`
`and Böhm, however, teach search methods that produce a meaningless “match” if
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`their neighbor searches are applied to the hashed extracted features as taught by
`
`Levy and Iggulden. Moreover, the proposed combinations produce the
`
`meaningless matches less efficiently than the exact match comparisons of Levy
`
`and Iggulden.
`
`The Petition fails to carry its burden under 37 C.F.R. § 42.104(b)(4) on each
`
`challenged independent claim (1, 5, 9, 13, 25, and 33) and thus fails to carry its
`
`burden on any of the challenged dependent claims (which all depend on the
`
`challenged independent claims). The Petition should be denied in its entirety.
`
`II. Overview of the ‘237 Patent.
`
`The ‘237 Patent involves a search that compares a given item to a reference
`
`database of potential matches. This Preliminary Response uses the term “work” to
`
`mean the item (e.g., a digital audio file) to be identified using the search. See ‘237,
`
`6:51-56. This Response uses “record” to mean one of the units in the reference
`
`database that the work may be compared to. See ‘237, 6:16-22. And this
`
`Response use “database” or “library” to mean the collection of all records. See
`
`‘237, 6:23-30.
`
`The independent claims of the ‘237 Patent include the following elements:
`
`[1] receiving or obtaining features extracted from a work;
`
`600885.2
`
`3
`
`
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`[2] identifying the work using the extracted features to perform a search of the
`
`database, where the search is:
`
` a “sub-linear time search” (claims 1 and 5);
`
` “an approximate nearest neighbor search” (claims 9 and 13);
`
` “a non-exhaustive search … to identify a near neighbor” (claims 25); or
`
` “a sublinear approximate nearest neighbor search” (claim 33); and
`
`[3] either (i) transmitting information about the identified work to the client
`
`device, or (ii) determining an action based on the identity of the work.
`
`The invention claimed in the ‘237 Patent includes two distinguishing features
`
`relevant to this Petition:
`
`Feature 1: Although the language varies among the claims, each claim
`
`requires that the “identifying” must be performed based on a search that has two
`
`components:
`
`(1) a sub-linear or non-exhaustive component (reflected in the underlined
`
`language):
`
` “sub-linear time search” (claims 1 and 5);
`
` “approximate nearest neighbor search” (claims 9 and 13);
`
` “non-exhaustive search … to identify a near neighbor” (claim 25); and
`
` “sublinear approximate nearest neighbor search” (claim 33).
`
`600885.2
`
`4
`
`
`
`(2) a neighbor component (reflected in the underlined language):
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
` a search that “identif[ies] a neighbor” (claims 1 and 5);
`
` “an approximate nearest neighbor search” (claims 9 and 13);
`
` “a nonexhuastive search … to identify a near neighbor” (claims 25); and
`
` “a sublinear approximate nearest neighbor search” (claim 33).
`
`Feature 2: The system must either determine an “action” based on the
`
`identification (claims 25 and 33); or transmit information about the identified
`
`media work to a “client device” (claims 1, 5, 9, and 13). It is not sufficient to
`
`simply to identify a match. Rather, an action must also be identified; or
`
`information about the identified work must be transmitted to the client device.
`
`III. Claim Constructions.
`
`A.
`
` “non-exhaustive search.”
`
`A “non-exhaustive” search is a search using an algorithm designed to locate
`
`a match without requiring the work to be compared to all records in the database.
`
`A “non-exhaustive” search uses an intelligent algorithm to narrow the database to
`
`only a subset of potential matches. See Ex. 1004, ¶12 (“algorithms that increased
`
`efficiency by intelligently searching only a subset of potential matches (i.e., ‘non-
`
`exhaustive’ algorithms)”). For example, if there are 100 records in a database, a
`
`non-exhaustive search could use an intelligent algorithm to exclude 75 records
`
`600885.2
`
`5
`
`
`
`from the search such that only 25 would be searched during the comparison
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`process. As the specification of the ‘237 Patent observes, these non-exhaustive
`
`“forms of matching include those based on clustering, kd-trees, vantage point trees
`
`and excluded middle vantage point forests” that do not systematically compare the
`
`work to each record. ‘237, 8:64-67.
`
`A “non-exhaustive” search can be contrasted with an “exhaustive” search.
`
`An exhaustive search systematically checks whether each record matches the work,
`
`“perhaps halting the search when the first match is found.” ‘237, 8:59-61. An
`
`exhaustive search is “a very general problem-solving technique that consists of
`
`systematically enumerating all possible candidates for the solution and checking
`
`whether each candidate satisfies the problem’s statement.” Ex. 2001.1 If there are
`
`100 records in the database, an “exhaustive” search does not narrow the potential
`
`matches but instead systematically compares the work with each record to
`
`determine if there is a match. Systematically comparing the work to be identified
`
`with each record rather than using intelligence to narrow the records to be
`
`compared is also referred to as using “brute force.” Ex. 1004, ¶44 (“a brute force
`
`search conducts a comparison of every item in a search database”); Ex. 2001.
`
`The Petition asserts that a “non-exhaustive search” should be construed as “a
`
`1
`Emphasis added throughout this Response, unless otherwise noted.
`
`600885.2
`
`6
`
`
`
`search that locates a match without conducting a brute force comparison of all
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`possible matches, and all data within all possible matches.” Pet. 5 (citing Ex.
`
`1004, ¶43). The “all data” clause (underlined above) in Petitioner’s proposed
`
`construction would improperly include as a “non-exhaustive search” any search
`
`that did not compare “all data” in each record, even if the search was a brute force
`
`comparison of each record in the database. As an illustrative example, assume the
`
`work to be identified “ABC” is compared with all records in a library, including
`
`record “DEF.” When comparing “ABC” with “DEF,” the algorithm determines
`
`that there is no match between “ABC” and “DEF” after just comparing the first
`
`letter of the work “A” with the first letter of the record “D.” If the algorithm does
`
`not unnecessarily compare the second and third letters, according to Petitioner, the
`
`search is not “exhaustive” even though every record is compared.
`
`To support its proposed construction, Petitioner relies solely on the
`
`conclusory assertions in its declaration (Pet. 5), and fails to present any objective
`
`evidence (e.g., cites to the ‘237 Patent specification and prosecution history,
`
`dictionary definitions, or treatises). Petitioner’s “all data” clause is improper
`
`because:
`
` it is inconsistent with how the non-exhaustive search concept is used in the
`
`‘237 Patent which describes a linear exhaustive search as one where the
`
`600885.2
`
`7
`
`
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`search compares the work to all “N entries,” not all data within all “N entries”
`
`(see ‘237, 21:7-40; 8:59-9:52); and
`
` it is not part of the ordinary meaning of “non-exhaustive search” (see Ex.
`
`2001).
`
`Petitioner’s declarant states that a non-exhaustive search is any search that is not a
`
`brute force search, and a “‘brute force’ search, in turn, is a search wherein a query
`
`is compared to every single portion of every single item in a database.” Ex. 1004
`
`¶43. The declarant provides no analysis or support for this conclusory assertion
`
`that is, therefore, insufficient to meet Petitioner’s burden. Smart Modular
`
`Technologies v. Netlist, Inc., IPR2014-01371, Paper 12 at 16 (PTAB March 13,
`
`2015) (“testimony is insufficient if it is merely conclusory” (quoting Schumer v.
`
`Lab. Computer Sys., Inc., 308 F.3d 1304, 1315 (Fed. Cir. 2002))).
`
`Moreover, objective sources confirm that an “exhaustive” or “brute-force”
`
`search systematically compares the work with each record in a database, not all
`
`data within each record, for example:
`
`In computer science, brute-force search or exhaustive search, also
`known as generate and test, is a very general problem-solving
`technique that consists of systematically enumerating all possible
`candidates for the solution and checking whether each candidate
`satisfies the problem’s statement.
`
`600885.2
`
`8
`
`
`
`Ex. 2001—each candidate is checked, not all data within each candidate.
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`In fact, Petitioner’s own declarant twice confirmed that a “non-exhaustive”
`
`search searches a subset of “potential matches,” not a subset of “all data within all
`
`potential matches”:
`
`Because neighbor searching is computationally intensive, content
`recognition schemes typically employed search algorithms that
`increased efficiency by intelligently searching only a subset of
`potential matches (i.e., ‘non-exhaustive’ algorithms).
`Ex. 1004 (Petitioner’s declaration), ¶12.
`
`to maximize search efficiency, persons skilled in the art routinely
`employed more efficient searches that did not conduct a comparison
`of every single item in a database, sometimes referred to as non-
`exhaustive searches.
`Ex. 1004 (Petitioner’s declaration), ¶43.
`
`As demonstrated below, the Petition fails under either the proper
`
`construction of “non-exhaustive” or under Petitioner’s improper construction that
`
`includes the “all data” clause.
`
`B.
`
`“sublinear.”
`
`A “sublinear search” (a type of non-exhaustive search) is “a search whose
`
`execution time scales with a less than linear relationship to the size of the data set
`
`to be searched, assuming computing power is held constant” where the “size of the
`
`600885.2
`
`9
`
`
`
`data set” is the number of potential matches in the data set (i.e., the “number of
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`entries in the search database”—Ex. 1004, ¶54).
`
`Petitioner asserts that a “sublinear search” is a “search whose execution time
`
`has a sublinear relationship to database size.” Pet. 6.
`
`As demonstrated below, the Petition fails under either construction.
`
`C.
`
` “neighbor” / “near neighbor.”
`
`Petitioner asserts that “identify a neighbor” and “identify a near neighbor”
`
`should all be construed to mean “identifying a close, but not necessarily exact,
`
`match.” Pet. 5-6 (also construing “neighbor” as “a close, but not necessarily
`
`exact, match”).
`
`Network-1 defines “identify a neighbor” and “identify a near neighbor” as
`
`“(i) as identifying a close, but not necessarily an exact or the closest, match of a
`
`feature, (ii) where “close” means a distance or difference that falls within a defined
`
`threshold, and (iii) where “feature” means a feature vector, compact electronic
`
`representation, or set of extracted features. The parties effectively agree on the
`
`first part of Network-1’s construction—“identifying a close, but not necessarily an
`
`exact or the closest, match.” The second part clarifies what constitutes a “close”
`
`600885.2
`
`10
`
`
`
`match as required by the ‘237 specification.2 The third part clarifies what is being
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`compared in the various claims of the ‘237 Patent and related patents.
`
`As demonstrated below, the Petition fails under either construction.
`
`D.
`
`“nearest neighbor.”
`
`Petitioner asserts that “near neighbor” and “nearest neighbor” have the exact
`
`same meaning. Pet. 5-6. This is incorrect. While as explained above, “neighbor”
`
`refers to a close, but not necessarily exact or the closest match, a “nearest neighbor,
`
`is the closest, but not necessarily exact match. ‘237, 9:14-15 (“A nearest neighbor
`
`search always finds the closest point to the query”).
`
`“When different words or phrases are used in separate claims, a difference in
`
`meaning is presumed.” Nystrom v. Trex Co., 424 F.3d 1136, 1143 (Fed. Cir.
`
`2005). “Near” and “nearest” have related but not identical meanings and Petitioner
`
`disregards this distinction. Moreover, Petitioner and its declarant provide no
`
`
`2
`The ‘237 specification describes “neighbor” searches as determining a match
`
`(or lack of match) based on a threshold. See e.g., ‘237, 6:38-40 (“If a match, or a
`
`match within a predetermined threshold is determined, then the associated work
`
`identifier 116 is read.”); id., 22:4-6 (“if the distance between the query and the
`
`nearest neighbor exceeds a threshold, then they are considered not to match.”)
`
`600885.2
`
`11
`
`
`
`evidence, support, or reasoning for its unsupported assertion that “near” and
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`“nearest” have identical meanings.
`
`E. “approximate nearest neighbor search”
`
`Petitioner fails to provide any construction of “approximate nearest neighbor
`
`search” and therefore fails to satisfy its burden as to the claims that include this
`
`phrase. The controlling statute mandates that a petition for IPR “may be
`
`considered only if ... the petition provides such other information as the Director
`
`may require by regulation.” 35 U.S.C. § 312. The controlling regulation states
`
`that “the petition must set forth ... how the challenged claim is to be construed.”
`
`37 C.F.R. § 42.104(b)(3). The Petition fails to identify any construction of the
`
`phrase “approximate nearest neighbor search.” Pet. 1-53. Accordingly, the
`
`Petition should be rejected at the threshold with respect to claims 9 and 13 (and
`
`claims dependent on these claims) that includes this phrase.
`
`The specification of the ‘237 Patent states that the claimed “approximate
`
`nearest neighbor search” is a sub-linear neighbor search, that does not always find
`
`the closest point to the query—i.e., does not always find the closest match:
`
`One example of a sub-linear time search is an approximate nearest
`neighbor search. A nearest neighbor search always finds the closest
`point to the query. An approximate nearest neighbor search does not
`always find the closest point to the query. For example, it might do so
`
`600885.2
`
`12
`
`
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`with some probability, or it might provide any point within some
`small distance of the closest point.
`‘237, 9:12-19.
`
`IV. Ground 1: Anticipated by Iwamura (Ex. 1012) fails.
`
`A. Overview of Iwamura.
`
`Iwamura discloses a “method to enable one to search for a song title when
`
`only its melody is known.” Iwamura, Abstract. “A remote music database with
`
`melody information is searched for the melody entered by the user, using for
`
`example, a peak or differential matching algorithm.” Iwamura, Abstract.
`
`The searching process disclosed in Iwamura is exhaustive rather than the
`
`claimed “non-exhaustive,” “sublinear,” or “approximate nearest” search. Iwamura
`
`discloses a searching algorithm that is designed to be more efficient than
`
`alternatives by matching up peak notes from the reference with the peak notes in
`
`the database when comparing the data. “Peak notes are also detected and marked
`
`when the data base is built.” Iwamura, 6:59-60. “A fast search is performed by
`
`using a peak or differential matching algorithm.” Iwamura, 12:1-2.
`
`While the individual comparisons of a work and a record in the library can
`
`be more efficient using the “peak note” approach disclosed in Iwamura (“search
`
`speed can be increased”), in doing so each record in the library is searched as part
`
`of the disclosed algorithm and “[t]he reference melody that gives the least
`
`600885.2
`
`13
`
`
`
`difference is returned as a search result.” Iwamura, 7:53-55.
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`B. Ground 1 fails because Iwamura does not disclose the claimed
`“non-exhaustive,” “sublinear,” and “approximate nearest” search
`limitations.
`
`Each independent claim of the ‘237 Patent includes a claimed a non-
`
`exhaustive, sublinear, or approximate nearest element:
`
` “sub-linear time search” (claims 1 and 5);
`
` “approximate nearest neighbor search” (claims 9 and 13);
`
` “nonexhaustive search … to identify a near neighbor” (claims 25); or
`
` “sublinear approximate nearest neighbor search” (claim 33).
`
`1.
`
`Iwamura does not disclose the claimed non-exhaustive,
`sublinear, or approximate nearest elements but rather
`teaches an exhaustive, linear search.
`
`Iwamura discloses an exhaustive, linear searching algorithm that is designed
`
`to be more efficient than alternatives by comparing peak notes from the reference
`
`with the peak notes in the database. Iwamura, 6:59-60, 12:1-2. While the
`
`individual comparisons of a work to a record in the library can be more efficient
`
`using this approach, Iwamura does not teach an algorithm that avoids an
`
`exhaustive search. Instead, each melody in the melody database is processed as
`
`part of the disclosed comparison and “[t]he reference melody that gives the least
`
`different is returned as a search result.” Iwamura, 7:53-55. Because the search
`
`600885.2
`
`14
`
`
`
`algorithm disclosed in Iwamura does not reduce the number of records to be
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`searched during a search but rather speeds up the comparison of the work to each
`
`record by matching peaks, the disclosed algorithm searches all records in the
`
`library and the computational time that the disclosed search takes to make such
`
`comparisons grows linearly with the number of records in the database.
`
`Accordingly, Iwamura teaches:
`
` an exhaustive search rather than the claimed “non-exhaustive” search,
`
`because it searches all records in the database;
`
` a linear search rather than the claimed “sublinear” search, because the
`
`computational time that it takes to perform a search grows linearly as new
`
`records are added to the database; and
`
` a nearest neighbor search, rather than the claimed “approximate nearest
`
`neighbor search” because Iwamura teaches that the melody that “gives the
`
`least difference is returned as a search result” (Iwamura, 7:52-55; ‘237, 9:14-
`
`17).
`
`The Petition (and corresponding declaration) fails to apply its constructions
`
`of “sublinear,” “non-exhaustive,” or “approximate nearest neighbor” (or any other
`
`constructions) to Iwamura. Accordingly, Petitioner fails to satisfy its burden of
`
`demonstrating that Iwamura anticipates because it fails to “map the construed
`
`600885.2
`
`15
`
`
`
`claim language to the teachings of the asserted prior art references.” International
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`Securities Exchange v. Chicago Board Options Exchange, IPR2014-00097, Paper
`
`12 at 26 (PTAB May 22, 2014).
`
`2.
`
`Response to Petitioner’s arguments.
`
`As support for the claimed non-exhaustive, sublinear, and approximate
`
`nearest search elements, Petitioner (and corresponding declaration) makes several
`
`assertions, none of which satisfy Petitioner’s burden of demonstrating that
`
`Iwamura discloses these claimed elements.
`
`“non-exhaustive search” (claim element 25(b)): Petitioner makes two
`
`assertions which are addressed in turn.
`
`Petitioner’s Assertion 1: “Iwamura further discloses non-exhaustive search
`
`algorithms using ‘peak notes’ (6:31-7:55), which ‘are approximately 20% of the
`
`total number of notes in a typical melody,’ meaning ‘search speed using peak notes
`
`is 20% of a brute force search’ (9:9-10).” Pet. 15.
`
`Under the proper construction of “non-exhaustive,” these passages from
`
`Iwamura do not disclose a non-exhaustive search because they do not state or
`
`suggest that all references in the music library are not used in the comparison.
`
`Rather, all reference melodies are compared and “[t]he reference melody that gives
`
`the least difference is returned as a search result.” Iwamura, 7:52-55.
`
`600885.2
`
`16
`
`
`
`Case No. IPR2015-00345
`U.S. Patent No. 8,205,237
`Patent Owner’s Preliminary Response
`
`
`Moreover, even using the “all data” clause in Petitioner’s improper
`
`construction—a non-exhaustive “search … locates a match without conducting a
`
`brute force comparison of … all data within all possible matches,”—the algorithm
`
`disclosed in Iwamura still falls under this definition because it compares “all data
`
`within all possible matches.” First, in implementing the comparison disclosed in
`
`Iwamura, every piece of data in every potential match must be analyzed to
`
`determine if it is a peak note—there is no way to identify the peak notes without
`
`analyzing every data point. See, e.g., Iwamura 7:1-4. Second, when comparing a
`
`song to be identified with each potential match, the peaks of the song to be
`
`identified are lined up with the peaks of the reference work to expedite the
`
`comparison: “In this manner, the entered melody is shifted to each peak in each
`
`reference melody and compared.” Iwamura, 7:52-55. But in doing so, this does
`
`not mean that only the peaks from the work to be identified are compared to the
`
`reference work. Rather, once the peaks are lined up, both the peaks and valleys
`
`(all data) is compared.
`
`Petitioner’s Assertion 2: “The search is further non-exhaustive because it
`
`can be accelerated