MICROSOFT TECHNOLOGY LICENSING, LLCDownload PDFPatent Trials and Appeals BoardOct 20, 20212020004071 (P.T.A.B. Oct. 20, 2021) Copy Citation UNITED STATES PATENT AND TRADEMARK OFFICE UNITED STATES DEPARTMENT OF COMMERCE United States Patent and Trademark Office Address: COMMISSIONER FOR PATENTS P.O. Box 1450 Alexandria, Virginia 22313-1450 www.uspto.gov APPLICATION NO. FILING DATE FIRST NAMED INVENTOR ATTORNEY DOCKET NO. CONFIRMATION NO. 15/186,223 06/17/2016 MICHAEL JOSEPH HOPCROFT 355652-US-NP/215340 8181 45809 7590 10/20/2021 SHOOK, HARDY & BACON L.L.P. (MICROSOFT TECHNOLOGY LICENSING, LLC) INTELLECTUAL PROPERTY DEPARTMENT 2555 GRAND BOULEVARD KANSAS CITY, MO 64108-2613 EXAMINER UDDIN, MOHAMMED R ART UNIT PAPER NUMBER 2167 NOTIFICATION DATE DELIVERY MODE 10/20/2021 ELECTRONIC Please find below and/or attached an Office communication concerning this application or proceeding. The time period for reply, if any, is set in the attached communication. Notice of the Office communication was sent electronically on above-indicated "Notification Date" to the following e-mail address(es): shbdocketing@shb.com usdocket@microsoft.com PTOL-90A (Rev. 04/07) UNITED STATES PATENT AND TRADEMARK OFFICE ____________ BEFORE THE PATENT TRIAL AND APPEAL BOARD ____________ Ex parte MICHAEL JOSEPH HOPCROFT, ROBERT LOVEJOY GOODWIN, FAN WANG, and UTKARSH JAIN __________ Appeal 2020-004071 Application 15/186,223 Technology Center 2100 ____________ Before JASON V. MORGAN, ERIC B. CHEN, and JAMES B. ARPIN, Administrative Patent Judges. ARPIN, Administrative Patent Judge. DECISION ON APPEAL Appellant1 appeals under 35 U.S.C. § 134(a) from the Examiner’s decision rejecting claims 1, 3–5, 9, 11–13, 15–17, 21–24, and 26–28, all of the pending claims. Final Act. 2.2 Claims 2, 6–8, 14, 18–20, and 25 are cancelled. We have jurisdiction under 35 U.S.C. § 6(b). 1 “Appellant” refers to “applicant” as defined in 37 C.F.R. § 1.42. Appellant identifies the real party-in-interest as Microsoft Technology Licensing, LLC. Appeal Br. 1. 2 In this Decision, we refer to Appellant’s Appeal Brief (“Appeal Br.,” filed December 5, 2019) and Reply Brief (“Reply Br.,” filed May 12, 2020); the Final Office Action (“Final Act.,” mailed July 17, 2019); the Advisory Action (“Adv. Act.,” mailed October 8, 2019); and the Examiner’s Answer (“Ans.,” mailed March 12, 2020); and the Specification (“Spec.,” filed June 17, 2016). Rather than repeat the Examiner’s findings and Appellant’s contentions in their entirety, we refer to these documents. Appeal 2020-004071 Application 15/186,223 2 We reverse. STATEMENT OF THE CASE The claimed methods, media, and systems relate to provid[ing] for a match fix-up stage that removes matching documents identified for a search query that don’t actually contain terms from the search query. A representation of each document (e.g., a forward index storing a list of terms for each document) is used to identify valid matching documents (i.e., documents containing terms from the search query) and invalid matching documents (i.e., documents that don’t contain terms from the search query). Any invalid matching documents are removed from further processing and ranking for the search query. Spec., Abstr. As noted above, claims 1, 3–5, 9, 11–13, 15–17, 21–24, and 26–28 are pending. Claims 1, 9, and 15 are independent. Appeal Br. 25 (claim 1), 26– 27 (claim 9), 27–28 (claim 15) (Claims App.). Claims 3–5 and 21–24 depend directly or indirectly from claim 1, claims 11–13 and 26–28 depend directly or indirectly from claim 9, and claims 16 and 17 depend directly from claim 15. Id. at 25–30. Claims 1, 9, and 15, reproduced below with disputed limitations emphasized, are illustrative. 1. A computer-implemented method, carried out by at least one server having one or more processors, the method comprising: employing a probabilistic matcher to identify, from a search index storing information regarding a plurality of documents, a set of matching documents by querying the search index with one or more terms, wherein the set of matching documents identified by the probabilistic matcher includes one or more valid matching documents having at least one of the one Appeal 2020-004071 Application 15/186,223 3 or more terms and one or more invalid matching documents not having any of the one or more terms; subsequent to identifying the set of matching documents based on the one or more terms, accessing a representation for each matching document of the set of matching documents, wherein the representation for each matching document includes a plurality of terms present within that matching document; comparing the terms present within each matching document to the one or more terms; identifying at least one invalid matching document from the one or more invalid matching documents by determining that the at least one invalid matching document does not include any of the one or more terms; and upon identifying the at least one invalid matching document, removing the at least one invalid matching document from the set of matching documents. Id. at 25 (emphasis added). 9. One or more computer storage media storing computer- useable instructions that, when used by one or more computing devices, cause the one or more computing devices to perform a method, the method comprising: querying a bit vector search index with one or more terms to identify a set of matching documents, the bit vector search index comprising a plurality of bit vectors, each bit vector comprising an array of bits that store information for a corresponding collection of terms contained in documents, wherein each bit in a given bit vector from the plurality of bit vectors indicates whether one or more documents corresponding with the bit include at least one term from the corresponding collection of terms for the given bit vector, wherein the set of matching documents includes one or more valid matching documents having at least one of the one or more terms and one or more invalid matching documents not having any of the one or more terms; Appeal 2020-004071 Application 15/186,223 4 receiving a forward index for each matching document of the set of matching documents, wherein the forward index includes one or more terms included in each matching document; using the forward index for each matching document of the set of matching documents, identifying at least one invalid matching document that does not include any of the one or more terms; and removing the at least one invalid matching document from the set of matching documents to create a filtered set of one or more matching documents. Id. at 26–27 (emphases added). 15. A computerized system comprising: one or more processors; and one or more computer storage media storing computer- useable instructions that, when used by the one or more processors, cause the system to: identify, from a search index, a set of matching documents by querying the search index with one or more terms, the set of matching documents including one or more valid matching documents having at least one of the one or more terms and one or more invalid matching documents not having any of the one or more terms; generate a preliminary ranking for each matching document from the set of matching documents; remove at least one matching document from the set of matching documents based on the preliminary rankings to provide a subset of matching documents; identify at least one invalid matching document in the subset of matching documents utilizing a forward index for each matching document of the subset of matching documents and remove the at least one invalid matching document to provide a filtered set of matching documents; generate a final ranking for at least a portion of the matching documents in the filtered set of matching a documents; and Appeal 2020-004071 Application 15/186,223 5 provide a set of search results based on the final rankings. Id. at 27–28 (emphases added). REFERENCES AND REJECTIONS The Examiner relies upon the following references: Name3 Reference Issued/Publ’d Filed Millett US 4,817,036 Mar. 28, 1989 Mar. 15, 1985 Schwenke US 2008/0177734 A1 July 24, 2008 Mar. 28, 2008 Wen US 2009/0012956 A1 Jan 8, 2009 Sept. 16, 2008 Grabarnik US 2009/0313202 A1 Dec. 17, 2009 June 13, 2008 The Examiner rejects: a. claims 1, 3–5, and 21–24 under 35 U.S.C. § 103 as obvious over the combined teachings of Wen and Schwenke (Final Act. 2–6); b. claims 9, 12, 13, and 26–28 under 35 U.S.C. § 103 as obvious over the combined teachings of Wen and Millett (id. at 7–10); c. claim 11 under 35 U.S.C. § 103 as obvious over the combined teachings of Wen, Millett, and Grabarnik (id. at 10–11); and d. claims 15–17 under 35 U.S.C. § 103 as obvious over the combined teachings of Wen and Grabarnik (id. at 11–14). We review the appealed rejections for error based upon the issues identified by Appellant, and in light of the contentions and evidence produced thereon. Ex parte Frye, 94 USPQ2d 1072, 1075 (BPAI 2010) (precedential). The Examiner and Appellant focus their findings and 3 All reference citations are to the first named inventor only. Appeal 2020-004071 Application 15/186,223 6 contentions, respectively, on claims 1, 9, and 15; so we do as well.4 See Appeal Br. 6, 14, 22; Ans. 3, 12, and 19. Because the reversals of the rejections of independent claims 1, 9, and 15 are dispositive, we do not address the merits of the rejections to the dependent claims further herein. We address the rejections below. ANALYSIS A. Obviousness Over Wen and Schwenke As noted above, the Examiner rejects claims 1, 3–5, and 21–24 as obvious over the combined teachings of Wen and Schwenke. Final Act. 2– 6. Appellant challenges the rejection of independent claim 1 separately. Appeal Br. 6–13. For the reasons given below, we are persuaded the Examiner errs, and we do not sustain this rejection. Claim 1 recites, employing a probabilistic matcher to identify, from a search index storing information regarding a plurality of documents, a set of matching documents by querying the search index with one or more terms, wherein the set of matching documents identified by the probabilistic matcher includes one or more valid matching documents having at least one of the one or more terms and one or more invalid matching documents not having any of the one or more terms. Appeal Br. 25 (Claims App.) (emphases added). Thus, we understand that, based on the plain language of claim 1, the recited set of matching documents includes both valid matching documents that include one or more of the query terms and invalid matching documents that include none of the 4 Appellant also challenges separately the rejection of claim 11, but our reversal of the rejection of independent claim 9 is dispositive of the rejection of claim 11. Appeal Br. 22; see Adv. Act. 3. Appeal 2020-004071 Application 15/186,223 7 query matching terms. See In re Morris, 127 F.3d 1048, 1054 (Fed. Cir. 1997) (“[T]he [Office] applies to the verbiage of the proposed claims the broadest reasonable meaning of the words in their ordinary usage as they would be understood by one of ordinary skill in the art, taking into account whatever enlightenment by way of definitions or otherwise that may be afforded by the written description contained in the applicant’s specification.”); see also Spec. ¶ 5 (“The technology described herein provides for a match fix-up stage to remove invalid matching documents returned from a bit vector search index.”) With respect to claim 1, the Examiner finds that Schwenke teaches or suggests this “employing” step. Final Act. 4 (citing Schwenke ¶ 50; Fig. 2 (reference numerals 201, 204)). In particular, Schwenke discloses that a query term “china” may result in exact matches as the highest probability matches and lower probability matches of “chair” and “chalk,” as partial matches due to the shared “ch” and non-matches, such as “table.” Schwenke ¶ 50, Fig. 1 (reference numerals 201–204); see Spec. ¶ 5 (describing false positives due to probabilistic approach). Thus, with reference to Schwenke, a valid matching set would include documents including “china,” and an invalid matching set would include documents including “chair,” “chalk,” and/or “table, but not “china.” Further, claim 1 recites, “subsequent to identifying the set of matching documents based on the one or more terms, accessing a representation for each matching document of the set of matching documents, wherein the representation for each matching document includes a plurality of terms present within that matching document.” Appeal Br. 25 (Claims App.) (emphases added). Thus, we understand that, once the set of matching Appeal 2020-004071 Application 15/186,223 8 documents has been identified, a “representation” of each valid and invalid matching document is accessed, and the “representation” includes “a plurality of terms present within that matching document.” See Spec. ¶ 5 (“Generally, a representation of each document is used to identify valid matching documents and invalid matching documents. The representation may be, for instance, a forward index that stores a list of terms for each document. Any invalid matching documents are removed such that they are not considered by the final ranker.” (emphasis added)). The plurality of “terms present” may include the query term, e.g., “china,” or other partial or non-matching terms, e.g., “chair,” “chalk,” and/or “table.” The Examiner finds Wen discloses, “The structured document retrieval process 400 continues to 416 in the XML structure, in which it searches for highest level and match, and then searches the next lower level for any matched terms.” and “The structured document retrieval process 400 continues to 418 in which those documents having the closest match are displayed.” Final Act. 3 (quoting Wen ¶¶ 72, 73). Referring to Wen’s paragraphs 42 and 49, the Examiner further finds: For example, when a user enter a query such as “history of china”, subsequent to identifying only documents with matching term “history of china” it also accessing representation of documents matching with other Chinese history such as “tang dynasty”, “ming dynasty”, etc. Representation of these documents includes plurality of queried term as claimed. Ans. 7 (citing Wen ¶¶ 42, 49). Based on these disclosures, the Examiner finds that Wen teaches or suggests the recited “accessing” step. For the reasons given below, we disagree with the Examiner. Appellant contends, “In accordance with claim 1, after a query is performed to identify a set of matching documents that includes valid and Appeal 2020-004071 Application 15/186,223 9 invalid matching documents, a representation of each of those matching documents is assessed.” Appeal Br. 9. In particular, Appellant contends that Wen teaches first ferreting or filtering out documents that do not contain the query term(s). Id. (quoting Wen ¶ 70 (“If the answer to decision 406 is [the particular structured document does not contains the index term(s)], then the structured document retrieval process 400 continues to 408 in which those documents that do not contain any of the search terms are ferreted (i.e., filtered) out.”)); see Wen, Fig. 7a (depicting filtering “invalid” matching documents (step 408) before performing operations on the structured documents (step 410)). Appellant contends Wen’s disclosure is in contrast to claim 1, in which a representation of each matching document from a set of matching documents is accessed in which the set of matching documents includes both: (1) one or more valid matching documents having at least one of the terms used to query; and (2) one or more invalid matching documents not having any of the terms used to query. In claim 1, a representation for both valid and invalid matching documents are accessed. Appeal Br. 9 (emphasis added); see Reply Br. 5 (“Based on the Examiner’s Answer, the Examiner appears to be arguing that the second element doesn’t require obtaining a representation for each valid matching document and each invalid matching document.”). This contention is consistent with our understanding of claim 1’s “accessing” step, and we agree with Appellant that the portions of Wen cited by the Examiner do not teach or suggest this limitation. Moreover, the Examiner’s findings regarding Wen’s paragraphs 42 and 49 do not persuade us otherwise. In paragraph 42, Wen discloses: The structured document retrieval system 100 uses the hierarchical indexing mechanism to assigns index terms that can Appeal 2020-004071 Application 15/186,223 10 precisely describe the inherent concept for each element while filtering out terms with too broad or too narrow semantics. Consider that a sample document about “China” contains a section “History”, which in turn also contains subsections such as “Tang dynasty”, “Ming dynasty” and “Qing dynasty”, etc. Selecting suitable index terms for the section “History” is challenging. “China” appears to be a relatively broad term that will provide too many results. “Tang”, “Ming”, and “Qing” by comparison appear to be narrow terms that may miss relevant documents. Terms such as “history” and “dynasty” appear to be good candidate terms. Wen ¶ 42; see id. ¶ 49 (“As such, the sections ‘China’ and ‘History’ form the index path from the query ‘history of China’; a user querying history of China returns this path.”). Thus, Wen describes documents that may contain the query term(s), as well as other terms, and possible disadvantages of filtering based on those other terms. We are not persuaded this discloses teaches or suggests, “accessing a representation” for each valid and invalid matching document. We are persuaded the Examiner errs in rejecting claim 1 as obvious over the combined teachings of Wen and Schwenke, and we do not sustain that rejection. Further, Appellant challenges the obviousness rejection of dependent claims 3–5 and 21–24 based only on its challenge to the rejection of their base claim, independent claim 1; and, therefore, we also do not sustain the rejection of those claims. B. Obviousness over Wen and Millett As noted above, the Examiner rejects claims 9, 12, 13, and 26–28 as obvious over the combined teachings of Wen and Millett. Final Act. 7–10. Appellant only challenges the rejection of independent claim 9 separately. Appeal 2020-004071 Application 15/186,223 11 Appeal Br. 13–22. For the reasons given below, we are persuaded the Examiner errs, and we do not sustain this rejection. Claim 9 recites, “wherein the set of matching documents includes one or more valid matching documents having at least one of the one or more terms and one or more invalid matching documents not having any of the one or more terms.” Appeal Br. 26 (Claims App.). Thus, as similarly noted above with respect to claim 1, we understand that, based on the plain language of claim 9, the recited set of matching documents includes both valid matching documents that include one or more of the query terms and invalid matching documents that include none of the query matching terms. See supra Section A. Although the Examiner relies on Schwenke to teach or suggest the identical limitation in claim 1,5 the Examiner finds Wen teaches or suggests this limitation of claim 9. Final Act. 7 (citing Wen ¶¶ 35, 36, 89, Figs. 3, 4, 7, 8). In particular, Wen discloses: Suppose a user who studies history of military operations intends to find out “what military aircrafts were used in Desert Storm”. An initial search may yield two articles Military Aircrafts and Gulf War as highly-ranking results both containing a portion with relevant content. The user has to scan each document (often quite long) to look for relevant content from the entire returned document. 5 The Examiner finds “Wen does not explicitly disclose . . . wherein the set of matching documents identified by the probabilistic matcher includes one or more valid matching documents and one or more invalid matching documents having at least one or more terms and one or more invalid matching document not having any of the one or more terms.” Final Act. 3–4 (Examiner’s emphasis); but see Adv. Act. 2–3 “(Wen teaches this concept and argued limitation in abstract, line 3-10, Para [0068].). Appeal 2020-004071 Application 15/186,223 12 Wen ¶ 89; see id. ¶ 36 (“Relevant document elements with arbitrary granularity can be obtained using a scalable element retrieve strategy by employing normal keyword-based queries.”). Based on these disclosures, the Examiner finds, “gulf war is the invalid matching document and military aircraft is the valid document for the search query what military aircrafts were used in Desert Storm.” Final Act. 7; Ans. 13. Appellant contends, instead of teaching a query that returns a set of matching documents that includes one or more valid matching documents (having a query term) and one or more invalid matching documents (not having any query term), Wen specifically discusses performing a query that removes all documents that do not have a query term (i.e., invalid matching documents). In particular, steps 406 and 408 of FIGS. 7 and 8 specifically indicate that documents that do not contain a term are filtered out. Additionally, [Wen] states: “those documents that do not contain any of the search terms are ferreted (i.e., filtered out).” As such, Wen only returns valid matching documents. Appeal Br. 16 (footnote omitted; emphasis added); see Reply Br. 9–10. We agree with Appellant. Wen discloses the query: “what military aircrafts were used in Desert Storm.” Wen ¶ 89. Based on the terms included in the query, two articles are retrieved each of which has “at least one of the one or more [query] terms.” Appeal Br. 26 (Claims App.). Wen may disclose the concept of searching for documents to identify valid and invalid matching documents and filtering out the invalid matching documents, but, given this filtering, the Examiner does not show Wen teaches the steps of “receiving a forward index for each matching document of the set of matching documents.” Final Act. 7–8 (citing Wen ¶ 71, Figs. 3, 4); see Appeal Br. 18–19 (missing “receiving” step); see also Appeal Br. 27 (Claims App.) (emphases added). Appeal 2020-004071 Application 15/186,223 13 Thus, Appellant contends each of Wen’s retrieved articles is a valid matching document; we agree with Appellant. Consequently, Appellant persuades us that the portions of Wen cited by the Examiner do not teach or suggest the disputed limitations. We are persuaded the Examiner errs in rejecting claim 9 as obvious over the combined teachings of Wen and Millett, and we do not sustain that rejection. Further, Appellant challenges the obviousness rejection of dependent claims 12, 13, and 26–28 based only on its challenge to the rejection of their base claim, independent claim 9; and, therefore, we also do not sustain the rejection of those claims. In addition, because we find the Examiner errs in rejecting independent claim 9 and because the Examiner does not contend that Grabarnik supplies the limitations of claim 9 that are missing from Wen and Millett, we also do not sustain the rejection of claim 11. See Final Act. 10–11; Appeal Br. 22; Ans. 18–19. C. Obviousness Over Wen and Grabarnik As noted above, the Examiner rejects claims 15–17 as obvious over the combined teachings of Wen and Grabarnik. Final Act. 11–14. Appellant only challenges the rejection of independent claim 15 separately. Appeal Br. 22–23. For the reasons given below, we are persuaded the Examiner errs, and we do not sustain this rejection. Claim 15 recites instructions causing one or more processors to identify, from a search index, a set of matching documents by querying the search index with one or more terms, the set of matching documents including one or more valid matching documents having at least one of the one or more terms and one or more invalid matching documents not having any of the one or more terms. Appeal 2020-004071 Application 15/186,223 14 Appeal Br. 28 (Claims App.) (emphasis added). As with the similar, disputed limitation in claim 9, the Examiner finds Wen teaches or suggests this disputed limitation. Final Act. 11–12 (citing Wen ¶¶ 35, 36, 89, Figs. 3, 4, 7, 8); see Ans. 20. Similar to its challenge to the rejection of claim 9, Appellant contends Wen does not teach this disputed limitation (see supra Section B) and Grabarnik does not cure the deficiencies of Wen (Appeal Br. 23). See Reply Br. 11 (“Instead, Wen discusses an approach ‘in which the search engine performs the search on all of the structured documents for the index terms’ to filter out any documents that don’t contain a query term. [Wen ¶¶ 69– 70.]”). For the same reasons given above with respect to claim 9, Appellant persuades us that the portions of Wen cited by the Examiner do not teach or suggest the disputed limitation. In addition, the Examiner finds Wen and Grabarnik teaches or suggests instructions causing one or more processors to “generate a preliminary ranking for each matching document from the set of matching documents; [and] remove at least one matching document from the set of matching documents based on the preliminary rankings to provide a subset of matching documents,” as recited in claim 15. Final Act. 12–13; see Appeal Br. 28 (Claims App.) (emphases added). However, because we are persuaded the Examiner fails to show that Wen teaches or suggests the set of matching documents including both valid and invalid documents, we also are persuaded the Examiner fails to show that Wen and Grabarnik teach or suggest, “generat[ing] a preliminary ranking” of the valid and invalid matching documents and “remov[ing] at least one matching document” from Appeal 2020-004071 Application 15/186,223 15 the set of valid and invalid matching documents based on the preliminary rankings. Appeal Br. 23. We are persuaded the Examiner errs in rejecting claim 15 as obvious over the combined teachings of Wen and Grabarnik, and we do not sustain that rejection. Further, Appellant challenges the obviousness rejection of dependent claims 16 and 17 based on its challenge to the rejection of their base claim, independent claim 15; and, therefore, we also do not sustain the rejection of those claims. DECISION 1. The Examiner errs in rejecting: a. claims 1, 3–5, and 21–24 as obvious over the combined teachings of Wen and Schwenke; b. claims 9, 12, 13, and 26–28 as obvious over the combined teachings of Wen and Millett; c. claim 11 as obvious over the combined teachings of Wen, Millett, and Grabarnik; and d. claims 15–17 as obvious over the combined teachings of Wen and Grabarnik. 2. Thus, on this record, claims 1, 3–5, 9, 11–13, 15–17, 21–24, and 26– 28 are not unpatentable. CONCLUSION We reverse the Examiner’s rejections of claims 1, 3–5, 9, 11–13, 15– 17, 21–24, and 26–28. Appeal 2020-004071 Application 15/186,223 16 In summary: Claims Rejected 35 U.S.C. § Reference(s)/Basis Affirmed Reversed 1, 3–5, 21–24 103 Wen, Schwenke 1, 3–5, 21– 24 9, 12, 13, 26–28 103 Wen, Millett 9, 12, 13, 26–28 11 103 Wen, Millett, Grabarnik 11 15–17 103 Wen, Grabarnik 15–17 Overall Outcome 1, 3–5, 9, 11–13, 15– 17, 21–24, 26–28 REVERSED Copy with citationCopy as parenthetical citation