Case 2:18-cv-00553-JRG-RSP Document 139 Filed 01/30/20 Page 1 of 12 PageID #: 6125
`
`IN THE UNITED STATES DISTRICT COURT
`FOR THE EASTERN DISTRICT OF TEXAS
`MARSHALL DIVISION
`
`
`
`UNILOC 2017 LLC,
`
`
`Plaintiff,
`
`
`v.
`
`GOOGLE LLC,
`
`
`
`Defendant.
`
`
`
`Civil Action No.: 2:18-cv-00553-JRG
`
`
`PATENT CASE
`
`JURY TRIAL DEMANDED
`
`
`UNILOC 2017’S REPLY BRIEF ON CLAIM CONSTRUCTION
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Page 1 of 12
`
`GOOGLE EXHIBIT 1006
`
`

`

`Case 2:18-cv-00553-JRG-RSP Document 139 Filed 01/30/20 Page 2 of 12 PageID #: 6126
`
`I. CONSTRUCTION OF DISPUTED TERMS.............................................................................. 1
`
`TABLE OF CONTENTS
`
`1.
`
`2.
`
`3.
`
`4.
`
`5.
`
`6.
`
`7.
`
`8.
`
`9.
`
`“keyfact” (all claims) ........................................................................................ 1
`
`“keyword” (claim 6) ......................................................................................... 2
`
`“extracting keywords without part-of-speech ambiguity” (claim 6) ................ 4
`
`“keyfact extracting step for analyzing a document collection and a user
`query, and extracting keywords without part-of-speech ambiguity from
`said document collection and said user query, and respectively extracting
`keyfacts of said document collection and said user query from said
`keywords” (claim 6) .......................................................................................... 4
`
`“keyfact indexing step for calculating the frequency of said keyfacts of
`said document collection and generating a keyfact list of said document
`collection for a keyfact index structure” (claim 6) ........................................... 6
`
`“keyfact retrieving step for receiving said keyfact of said user query and
`said keyfacts of said document collection” (claim 6) ....................................... 7
`
`“keyfact retrieving step for . . . defining a keyfact retrieval model in
`consideration of weigh factors according to a keyfact pattern” (claim 6) ........ 7
`
`“keyfact retrieving step for . . . generating a retrieval result” (claim 6) ........... 7
`
`Order of steps of claim 6 of the ’908 patent (claim 6) ...................................... 8
`
`10.
`
`Order of steps of claim 11 of the ’908 patent (claim 11) .................................. 8
`
`i
`
`Page 2 of 12
`
`

`

`Case 2:18-cv-00553-JRG-RSP Document 139 Filed 01/30/20 Page 3 of 12 PageID #: 6127
`
`I. CONSTRUCTION OF DISPUTED TERMS
`
`1.
`
`“keyfact” (all claims)
`
`Google’s attorney argument concerning the term “keyfact” is based on a “gotcha.”
`
`Google contends that a construction of “keyfact” should include the word “important” and then
`
`“gotcha,” inclusion of that word would render the claim indefinite. Dkt. 132 at 2-5. Google bases
`
`its argument on a single snippet from the specification of the ’908 patent using the word
`
`“important,” but fails to show how a person of ordinary skill in the art would understand the term
`
`“keyfact” from the context of the specification of the ’908 patent. And despite Google’s burden
`
`to demonstrate indefiniteness by clear and convincing evidence, Google provides no evidence or
`
`testimony that a person of ordinary skill in the art would be unable to understand what a
`
`“keyfact” is after reading the ’908 patent.
`
`Google is also incorrect in limiting the construction of a “keyfact” to documents.
`
`Demonstrating Google’s error, Google exclusively relies on the sentence, “A keyfact means an
`
`important fact contained in sentences which constitute a document.” Dkt. 134 at 2 (citing ’908
`
`patent, 1:15-16). But the ’908 patent makes clear that a “keyfact” can be extracted from a user
`
`query as well as a “document.” Accordingly, Google’s “gotcha” construction cannot be correct
`
`as it improperly relies on a sentence from the specification that excludes exemplary disclosed
`
`“keyfacts,” in particular those extracted from user queries.
`
`Indeed, a “keyfact,” as that term is used in the ’908 patent, applies to both analyzing a
`
`document collection as well as user queries into the document collection. See ’908 patent,
`
`Abstract (“a document collection and a query”). A particular keyfact in a document becomes
`
`important in the context of the entire system of analyzing a document collection, parsing queries,
`
`and generating results. A “keyfact” does not, however, possess an intrinsic “importance.” When
`
`considering the retrieval system described in the ’908 patent, Uniloc properly construes a
`
`“keyfact” as a “fact contained in sentences” of a document in a document collection or a user
`
`query. A particular “keyfact” in that collection becomes “important” when a user constructs a
`
`query to search for the document and that query contains a similar “keyfact.” ’908 patent, Fig. 6;
`
`1
`
`Page 3 of 12
`
`

`

`Case 2:18-cv-00553-JRG-RSP Document 139 Filed 01/30/20 Page 4 of 12 PageID #: 6128
`
`8:25-34. Further, the frequency of how many times a “keyfact” can be found in a document can
`
`show its relative importance. ’908 patent, Table 2.
`
`Even the prior work of the inventors concerning “keyfacts” (cited by Google as extrinsic
`
`evidence, Ex. A of Dkt. 132) shows that the word “important” is not part of how a person
`
`ordinary skill would understand a “keyfact.” The inventor’s other explanation of the concept of
`
`a “keyfact” did not rely on the notion of “important.” See Ex. A at Abstract (“We shall show
`
`how to extract a fact from a document using an extended concept of keyword, called keyfact
`
`which can contain syntactic patterns and semantic information.”). Given the full context of the
`
`’908 patent and the prior use of the term “keyfact” in Ex. A, there is simply no reason to think a
`
`person of ordinary skill would find the term “keyfact” indefinite.
`
`Google attempts to criticize Uniloc’s construction of “keyfact” as overly broad, but
`
`Uniloc does not construe “keyfact” as a “fact.” See Dkt. 132 at 7. Uniloc’s construction—a “fact
`
`contained in sentences”—relies on the semantic nature of a sentence to distinguish from bare
`
`“facts,” and Uniloc’s construction should be read in the context of the remaining elements of the
`
`claims to perform the claimed method. Incidentally, if Google contends that “key” means
`
`“important,” then that notion runs contrary to Google’s proposed construction of “keyword,”
`
`which does not rely on the concept of “important.”
`
`2.
`
`“keyword” (claim 6)
`
`Like its incorrect construction of “keyfact,” Google’s construction of “keyword” also
`
`fails to recognize that a “keyfact” can be based on words in a document or words within a query
`
`used to retrieve documents. Dkt. 128 at 8. Google’s recent removal of the words “within a
`
`query” from its construction (Dkt. 134 at 8 n.5) does nothing to resolve the problem that
`
`keywords can be found in documents themselves. See, e.g., ’908 patent, 5:15-18; 8:26-29. Claim
`
`6, in fact, recites “extracting keywords … from said document collection and said user query.”
`
`(emphasis added). Accordingly, Google’s attempt to define a “keyword” as “used to retrieve
`
`documents” should fail.
`
`2
`
`Page 4 of 12
`
`

`

`Case 2:18-cv-00553-JRG-RSP Document 139 Filed 01/30/20 Page 5 of 12 PageID #: 6129
`
`Turning to the parties’ dispute of whether a “keyword” is limited to a “noun,” Google
`
`contends that the specification provides a clear definition of “keyword.” More specifically,
`
`Google contends that the ’908 patent’s background refers to “keywords, which are nouns” and
`
`therefore that the patentee acted as a lexicographer to define the word “keyword.” Dkt. 134 at 8.
`
`But this discussion in the background reflects no clear intent to restrict “keywords” to nouns, and
`
`the specification, not the background, shows that “keywords” of the ’908 patent are in fact not
`
`restricted to “nouns.”
`
`Claim 6 of the ’908 patent requires “extracting keywords without part-of-speech
`
`ambiguity.” As discussed in the ’908 patent regarding a document in a document collection:
`
` A
`
` document is supplied at stage 31 and morphological analysis is
`performed at stage 32. A sentence in the document is divided into
`words and the morphological analysis is performed with
`dictionaries 36 at stage 32. The morphological variation is
`considered in order to recover prototypes. The dictionaries 36
`include a noun dictionary, a verb dictionary, an adjective
`dictionary, an adverb dictionary, a preposition dictionary, a
`conjunction dictionary, and a stop-word lexicon. In some cases, a
`part-of-speech of a word is determined by rules without
`dictionaries.
`
`’908 patent, 5:19-28. It could not be clearer from this passage that the “words” extracted are not
`
`only nouns. Moreover, those extracted words are used to form parts of the keyfact. ’908 patent,
`
`5:29-36. Indeed, while the specification states an object is a “noun or compound nouns
`
`represented by a keyword,” other portions of the keyfact are also represented by keywords. See,
`
`e.g., ’908 patent, 6:33-37 (associating tag MP1 with adjective keyword “fast” and tag MP2 with
`
`verb keyword “distributed”). Google’s brief focuses on how the preferred embodiment
`
`constructs keyfacts using the key tag KEY and appears to contend that all KEY key tags are
`
`nouns or compound nouns. Dkt. 134 at 10-11. But Google’s analysis fails because the preferred
`
`embodiment uses other tags, and other parts of speech, and those other tags are also represented
`
`by keywords. Google’s construction is thus improperly narrow.
`
`
`
`3
`
`Page 5 of 12
`
`

`

`Case 2:18-cv-00553-JRG-RSP Document 139 Filed 01/30/20 Page 6 of 12 PageID #: 6130
`
`3.
`
`“extracting keywords without part-of-speech ambiguity” (claim 6)
`
`Google’s argument on this term completely refutes its proposed construction of
`
`“keyword,” as it demonstrates that “keyfacts” are extracted using “keywords,” and “keywords”
`
`can have part-of-speech ambiguity. See Dkt. 134 at 13 (acknowledging “keyfacts are respectively
`
`extracted from those previously extracted keywords from the document collection and user
`
`query.”)
`
`Google’s construction of “extracting keywords without part-of-speech ambiguity” also
`
`cannot be correct, as it suggests that words found in a document (or a query) are tagged and then
`
`only those found with a single tag are extracted from the document or query. Rather, as detailed
`
`by Uniloc in its opening brief, words in a document (or query) are extracted. Then, those words
`
`are processed to resolve part-of-speech ambiguity. This does not mean, as Google contends, that
`
`“only” certain words are extracted. Dkt. 134 at 15 (contending, “Only keywords from the
`
`document collection and user query that do not have any part-of-speech ambiguity are
`
`extracted.”). Rather, as explained in the ’908 patent, a “morphological analysis” analyzes the
`
`words of the document (or query) to resolve ambiguities in their part of speech. This can be done
`
`using a “part of speech tag in dictionaries” (5:29-30) or “a part-of-speech of a word is
`
`determined by rules without dictionaries” (5:27-28). Google’s construction should be rejected.
`
`4.
`
`“keyfact extracting step for analyzing a document collection and a user
`query, and extracting keywords without part-of-speech ambiguity from said
`document collection and said user query, and respectively extracting keyfacts
`of said document collection and said user query from said keywords” (claim
`6)
`
`Google fails to show that the “keyfact extracting step for analyzing a document collection
`
`and a user query” invokes 35 U.S.C. § 112(f). In particular, Google parses this element in a
`
`manner that avoids reading “extracting keywords…” and “extracting keyfacts…” as acts
`
`performing the function of “analyzing a document collection and a user query.” This is error, as
`
`the acts of “keyfact extracting” are part of the analysis of the document collection or user query
`
`to extract keyfacts.
`
`4
`
`Page 6 of 12
`
`

`

`Case 2:18-cv-00553-JRG-RSP Document 139 Filed 01/30/20 Page 7 of 12 PageID #: 6131
`
`Google contends that col. 4, lines 40-46 of the ’908 patent provides that “‘analyzing a
`
`document collection and a user query’ is merely a predicate step before keywords and keyfacts
`
`may be extracted,” but to the contrary that passage supports Uniloc’s position that extracting
`
`keywords and extracting keyfacts are acts that are part of analyzing a document collection
`
`(emphasis added):
`Now, the keyfact-based text retrieval system in accordance with
`the present invention is explained with reference to FIG. 1. Once a
`document collection 14 or a query 15 is given, the keyfact
`extraction device 11 extracts words without ambiguity by
`performing morphological analysis and tagging. The keyfact
`generation rule is applied to the words and then the keyfacts are
`extracted.
`
`Moreover, claim 6 also states how the acts are performed: keywords are extracted “without part-
`
`of-speech ambiguity” and keyfacts are extracted “from said keywords.” With such keyfacts
`
`extracted, a document collection or user query is analyzed. No more is required to avoid
`
`application of 35 U.S.C. § 112(f) for this element.
`
`Google’s allegations of acts corresponding to the alleged functions demonstrate that
`
`Google is adding acts unnecessary to perform the alleged functions. Even under a § 112(f)
`
`construction, this would be improper, as § 112(f) does not “permit incorporation of structure
`
`from the written description beyond that necessary to perform the claimed function.” Micro
`
`Chemical, Inc. v. Great Plains Chemical Co., Inc., 194 F.3d 1250, 1258 (Fed. Cir. 1999) (dealing
`
`with “means” elements). For example, for “analyzing a document collection and a user query,”
`
`Google contends “the corresponding acts in the specification describe dividing a sentence into
`
`words and performing morphological analysis with dictionaries, where ‘morphological variation
`
`is considered in order to recover prototypes.’” Dkt. 134 at 17. Putting aside the overlap of these
`
`acts in the preferred embodiment with the act of “extracting keywords,” use of dictionaries and
`
`“prototypes” is not necessary to perform the claimed analysis. In fact, the ’908 patent states that
`
`the use of dictionaries is optional: “In some cases, a part-of-speech of a word is determined by
`
`rules without dictionaries.” ’908 patent, 5:27-28. With regard to the act of “extracting
`
`5
`
`Page 7 of 12
`
`

`

`Case 2:18-cv-00553-JRG-RSP Document 139 Filed 01/30/20 Page 8 of 12 PageID #: 6132
`
`keywords,” claim 6 already recites “without part-of-speech ambiguity,” and there is no need to
`
`recite the additional act of “tagging,” as contended by Google. Similarly, the act of “keyfact
`
`extraction” requires use of the extracted keywords, and there is no function recited in claim 6
`
`that would require the acts using a “keyfact pattern rule” or keyfacts in a particular format. Cf.
`
`Dkt. 134 at 23 (improperly contending claim 6 requires “keyfact terms are represented in the
`
`form ‘[object, property]’ and describe how possible keyfact lists are generated from each keyfact
`
`pattern based on the keyfact generation rule”).
`
`Accordingly, Google’s attempt to add the acts of the preferred embodiment into claim 6
`
`should be rejected.
`
`5.
`
`“keyfact indexing step for calculating the frequency of said keyfacts of said
`document collection and generating a keyfact list of said document collection
`for a keyfact index structure” (claim 6)
`
`Google fails to show that the “keyfact indexing” element invokes 35 U.S.C. § 112(f).
`
`Google first presents a circular argument assuming the element is in a “step-plus-function”
`
`format and then requiring there be a function “after the … ‘steps for’ language.” Dkt. 134 at 19.
`
`But Uniloc’s presentation demonstrated that there is no “function” “after the … ‘steps for’
`
`language. The “keyfact indexing” step includes only acts after the “steps for” language, which
`
`demonstrates that the element is not in “step-plus-function” format. See Wenger Mfg., Inc. v.
`
`Coating Mach. Sys., Inc., 239 F.3d 1225, 1237 (Fed. Cir. 2001); York Products, Inc. v. Central
`
`Tractor Farm & Family Center, 99 F.3d 1568, 1574 (Fed. Cir. 1996) (“Without an identified
`
`function, the term ‘means' in this claim cannot invoke 35 U.S.C. § 112, ¶ 6.”). And there is no
`
`ambiguity in Uniloc’s position regarding the “keyfact extracting” and “keyfact indexing”
`
`elements. The “keyfact extracting” element does not invoke § 112(f) because acts later recited in
`
`the element perform the function of “analyzing,” see Rodime PLC v. Seagate Tech., Inc., 174
`
`F.3d 1294, 1302 (Fed. Cir. 1999) (“[E]ven if the claim element specifies a function, if it also
`
`recites sufficient structure or material for performing that function, § 112, ¶ 6 does not apply”),
`
`and “keyfact indexing” does not invoke § 112(f) because it recites acts to perform “keyfact
`
`indexing.”
`
`6
`
`Page 8 of 12
`
`

`

`Case 2:18-cv-00553-JRG-RSP Document 139 Filed 01/30/20 Page 9 of 12 PageID #: 6133
`
`Google’s allegations of acts corresponding to the alleged functions demonstrate that
`
`Google is adding acts unnecessary to perform the alleged functions. Even under a § 112(f)
`
`construction, this would be improper, as § 112(f) does not “permit incorporation of structure
`
`from the written description beyond that necessary to perform the claimed function.” Micro
`
`Chemical, 194 F.3d at 1258 (dealing with “means” elements). In fact, for this element, Google
`
`does not actually contend that the claimed “keyfact indexing” is not achieved by “calculating”
`
`and “generating.” Rather, Google seeks to add, for the “calculating the frequency of said
`
`keyfacts,” that “[f]or each document, the keyfact frequency (tf) and document frequency of the
`
`keyfact (df) are calculating [sic] in order to obtain the frequency information of the keyfacts.”
`
`Dkt. 134 at 20. Thus, for the “calculating the frequency,” Google requires both (tf) and (df),
`
`which is not required by claim 6. Similarly, for the act of “generating a keyfact list of said
`
`document collection for a keyfact index structure,” Google attempts to saddle the claim with
`
`additional acts involving (tf) and (df), no less than three “tables,” rather than simply the “list”
`
`required by the claim. Dkt. 134 at 20-21.
`
`Accordingly, Google’s attempt to add the acts of the preferred embodiment into claim 6
`
`should be rejected.
`
`6.
`
`7.
`
`“keyfact retrieving step for receiving said keyfact of said user query and said
`keyfacts of said document collection” (claim 6)
`
`“keyfact retrieving step for . . . defining a keyfact retrieval model in
`consideration of weigh factors according to a keyfact pattern” (claim 6)
`
`8.
`
`“keyfact retrieving step for . . . generating a retrieval result” (claim 6)
`
`Google uses the same circular reasoning for the “keyfact retrieving” elements as it used
`
`for the “keyfact indexing” element. Dkt. 134 at 22-23, 24. Each of terms (6)-(8) recite the acts of
`
`“keyfact retrieving” rather than functions. Accordingly, this element is not in “step-plus-
`
`function” format, and § 112(f) does not apply. See Wenger Mfg., Inc., 239 F.3d at 1237; York
`
`Products, Inc., 99 F.3d at 1574; Rodime PLC, 174 F.3d at 1302.
`
`As Google did for other elements, Google similarly adds acts not contemplated by claim
`
`6. Even under a § 112(f) construction, this would be improper, as § 112(f) does not “permit
`
`7
`
`Page 9 of 12
`
`

`

`Case 2:18-cv-00553-JRG-RSP Document 139 Filed 01/30/20 Page 10 of 12 PageID #: 6134
`
`incorporation of structure from the written description beyond that necessary to perform the
`
`claimed function.” Micro Chemical, 194 F.3d at 1258 (dealing with “means” elements). For
`
`example, Google acknowledges, but does not resolve, that some acts overlap with others. Dkt.
`
`134 at 23-24. Also, for the act of “defining a keyfact retrieval model in consideration of weigh
`
`factors according to a keyfact pattern,” Google adds the acts of how “weigh factors” may be
`
`determined. Dkt. 134 at 24-25 (requiring a “document vector and query vector based on the
`
`received keyfacts” and a equation set forth in the preferred embodiment). Also, for the act of
`
`“generating a retrieval result,” Google seems to require a “vector space model” and generating
`
`an “order.” Neither of these acts are necessary to generate a retrieval result, as claimed.
`
`Accordingly, Google’s attempt to add the acts of the preferred embodiment into claim 6
`
`should be rejected.
`
`9.
`
`Order of steps of claim 6 of the ’908 patent (claim 6)
`
`Google’s error in requiring a blanket order-of-steps construction lies in the fact that the
`
`steps of the claimed method contain multiple acts that can be performed in a parallel or
`
`overlapping manner. In particular, Google does not account for the separate recitations directed
`
`to “a document collection” and a “user query.” A “keyfact extracting” step applied to a user
`
`query could be performed before or after a “keyfact extracting” step as applied to a document
`
`collection (or before or after the “keyfact indexing” step, or before or after a “keyfact retrieving”
`
`steps as applied to a document collection for that matter). The “keyfact indexing” step does not
`
`even recite a “user query,” and Google does not contend otherwise. For example, a document
`
`collection could be indexed prior to a “user query” initiating a “keyfact extracting” step for the
`
`“user query.” Thus, it makes no sense to say that the steps “must occur in the order recited in the
`
`claims.”
`
`10. Order of steps of claim 11 of the ’908 patent (claim 11)
`
`As noted above, Google insufficiently shows that the steps of claim 6 should be
`
`interpreted as a discrete series of steps. Claim 11, like claim 6, does not require such an order of
`
`steps.
`
`8
`
`Page 10 of 12
`
`

`

`Case 2:18-cv-00553-JRG-RSP Document 139 Filed 01/30/20 Page 11 of 12 PageID #: 6135
`
`The disputed claim terms should be construed as proposed by Uniloc.
`
`
`Dated: January 30, 2020
`
`
`
`
`
`Respectfully submitted,
`
`
`
`
`
`/s/ James L. Ethridge
`James L. Etheridge, TX Bar No. 24059147
`Ryan S. Loveless, TX Bar No. 24036997
`Brett A. Mangrum, TX Bar No. 24065671
`Travis L. Richins, TX Bar No. 24061296
`Jeff Huang (Admitted E.D. Tex.)
`Etheridge Law Group, PLLC
`2600 E. Southlake Blvd., Suite 120 / 324
`Southlake, TX 76092
`Tel.: (817) 470-7249
`Fax: (817) 887-5950
`Jim@EtheridgeLaw.com
`Ryan@EtheridgeLaw.com
`Brett@EtheridgeLaw.com
`Travis@EtheridgeLaw.com
`Jeff@EtheridgeLaw.com
`
`
`
`
`9
`
`Page 11 of 12
`
`

`

`Case 2:18-cv-00553-JRG-RSP Document 139 Filed 01/30/20 Page 12 of 12 PageID #: 6136
`
`
`
`CERTIFICATE OF SERVICE
`
`I hereby certify that all counsel of record who are deemed to have consented to electronic
`
`service are being served with a copy of this document via the Court’s CM/ECF system per Local
`
`Rule CV-5(a)(3).
`
`
`
`DATED: January 30, 2020
`
`
`
`
`
`
`
`/s/ James L. Etheridge
`James L. Etheridge
`
`
`
`
`10
`
`Page 12 of 12
`
`