Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Detailed Action
In amendments dated 4/22/26, Applicant amended claims 1 and 10, canceled claims 4-5 and 13-14, and added no new claims. Claims 1-3, 6-12, and 15-19 are presented for examination.
Rejections under 35 U.S.C. 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-3, 6-12, and 15-19 are rejected under 35 U.S.C. 101 because the claimed invention is directed to mental processes without significantly more. Independent claims 1 and 10 each recites automatically detecting a line of business associated with a user based upon a device used by the user to send a text query or automatically determining the line of business associated with the user based on a product or a text query received from a device associated with a user account; generating, using an encoder configured to encode a query embedding in a vector space from the text query as a real-valued vector representation of semantic meaning of the text query such that words or sentences from the query that are similar in sematic meaning are closer to each other in the vector space; adding the documents to a reverse index by breaking down unstructured text of the documents into individual sentences while preserving semantic relationships between words and sentences using customized metadata; extracting titles from the documents or from metadata associated with the documents or generating titles from contents of the individual sentences, extracting entity tags from the documents using a custom analyzer that isolates entity-specific information such that alphanumeric labels with special characters are not inadvertently separated, parsed, or removed during preprocessing, and generating title embeddings for the titles and sentence embeddings for the individual sentences; scoring entries in the reverse index using a hybrid scoring function, wherein the hybrid scoring function is used to generate scores, each generated score based on five individual scores that comprise keyword matching between (i) the text query and the titles, (ii) the text query and the individual sentences, and (iii) the line of business associated with the user and the entity tags, and cosine similarity between (i) the query embedding and the title embeddings and (ii) the query embedding and the sentence embeddings, and wherein the five individual scores of the hybrid scoring function are computed in parallel using an index structure; and ranking the generated scores for the individual sentences in the document database. Detecting or determining a line of business is evaluating and a mental process, generating a query embedding using an encoder is recited broadly and is a mental process accomplishable in the human mind or on paper. Examiner notes that generating embeddings using an encoder is applying said encoder which is invokes the computer as a tool per MPEP 2106.05(f)(2) and is not significantly more than an abstract idea per Recentive Analytics v. Fox Broadcasting Corp. (134 F.4th 1205, 2025 U.S.P.Q.2d 628). Extracting or generating titles or entity tags or title embeddings, scoring entries, and ranking the generated scores are each recited broadly and a mental process accomplishable in the human mind or on paper. Each claim recited additional elements of receiving documents in the document database and receiving the text query from the user via the device, which are each data gathering steps and insignificant extra-solution activity; storing the titles, the title embeddings, the individual sentences, the sentence embeddings, and the entity tags in the reverse index, which is insignificant extra-solution activity; and displaying an individual sentence associated with a top score, from among the ranked scores, to the user, an output step and also insignificant extra-solution activity. Claim 10 also recites at least on processor comprising a document database and software components and a computer-readable memory device which are each generic components of a computer system. Examiner notes printed specification paragraph 0030 states “keyword similarity search and embedding similarity search methods are effective for small search spaces; however, the accuracy begins to decrease with as the search space increases in size.” Paragraph 0030 then says “optimization techniques may be used to scale the search methods to larger search spaces” and “optimization techniques include, for example, multi-stage search, metadata optimization, and ranking.” Paragraph 0031 describe the multi-stage search technique for reducing the search space to a subset of relevant documents and then searching those documents; paragraph 0032 describes the metadata technique also reduces a search space using data related to entities in the documents or the end user and then searches relevant documents; and per paragraph 0033 “ranking is used to prioritize search results based on their past utility” and this technique “boosts search results that have been previously relevant for similar queries” and “this method also requires the collection of direct or indirect feedback to assess the utility of search results.” The claims recite extracting metadata from documents but do not recite reducing a search space to a subset of documents using metadata or a multi-stage process, or using ranking to require feedback and prioritize search results based on their past utility. Thus the claim steps still do not recite a particular improvement in any technology or function of a computer per MPEP 2106.04(d) and do not recite any unconventional steps in the invention per MPEP 2106.05(a). Therefore, the recited mental processes are not integrated into a practical application. Taking the claims as a whole, the data input steps and output step are recited broadly and amount to sending and receiving data across a network per specification paragraphs 0094, 0100, 0107, figure 7 706 and figure 6 606, which is routine and conventional activity per the list of said activities in MPEP 2106.05(d) part II. The storing step is also routine and conventional activity per the list of such activities in MPEP 2106.05(d) part II. The at least one processor device and at least one computer-readable memory are still each generic components of a computer. Therefore, the claims do not include additional elements that are sufficient to amount to significantly more than the recited mental processes.
Claims 2 and 11 each recites wherein the hybrid scoring function has the form: Hybrid Score=(1+Keyword Match Score) x (1+Cosine Similarity Score), which is a mathematical formula but is applied to generate scores and is therefore insufficient to amount to significantly more than the recited mental processes. Claims 3 and 12 each recites wherein the hybrid scoring function has the form: Hybrid Score = (1 + k1) X (1 + k2) X (1 + k3) X (1 + sl) X (1 + s2), and wherein k1 represents a score for a first keyword matching between the text query and the titles; k2 represents a score for a second keyword matching between the text query and the sentences; k3 represents a score for a third keyword matching between the user line of business and the entity tags;s1 denotes a first search score based on embeddings, which measures cosine similarity between the query embedding and the title embeddings; and s2 denotes a second search score based on embeddings, which measures cosine similarity between the query embedding and the sentence embeddings, and the hybrid scoring function is a mathematical formula but is applied to generate scores and is therefore insufficient to amount to significantly more than the recited mental processes.
Claims 6 and 15 each recites wherein the user line of business and the entity tags comprise one or more of: product codes or names, product categories, product types, error codes, lines of business, model numbers, serial numbers, and entity brand names, which are each data and thus mental process accomplishable in the human mind or on paper. Claims 7 recites further comprising: dividing the reverse index into shards, and dividing a data structure is a mental process accomplishable in the human mind or on paper; storing groups of the shards at two or more locations, and storing shards is storing data and is routine and conventional activity per the list of said activities in MPEP 2106.05(d) part II; and scoring entries in multiple reverse index shards simultaneously, and scoring entries is recited broadly and a mental process accomplishable in the human mind or on paper. Claims 8 and 18 each recites creating replicas of the shards; and storing groups of the replicas at the two or more locations, which are both storing data and routine and conventional activity per the list of said activities in MPEP 2106.05(d) part II.
Claims 9 and 19 each recites providing the user with a link to a source document for the sentence associated with the top score, and providing data to a user is recited broadly and amounts to sending data across a network which is routine and conventional activity per the list of said activities in MPEP 2106.05(d) part II. Claim 16 recites a plurality of the computer-readable memory devices, wherein shards of the index are stored across different ones of the memory devices, and storing data is a routine and conventional activity per the list of said activities in MPEP 2106.05(d) part II. Claim 17 recites wherein the search component is further configured to score the entries in separate index shards simultaneously, and scoring entries is recited broadly and a mental process accomplishable in the human mind or on paper.
Rejections under 35 U.S.C. 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1, 10, and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Kotaru (US 20250259096) in view of Pearman et al (US 20240184789), hereafter Pearman.
With respect to claims 1 and 10, Kotaru teaches:
receiving documents in the document database (paragraph 0032 example documents of technical specifications in a database);
automatically detecting a line of business associated with a user based upon a device used by the user to send a text query or automatically determining the line of business associated with the user based on a product or a text query received from a device associated with a user account (paragraph 0005 line of business detected from technical documents, paragraph 0020 example of users in an engineering team);
receiving the text query from the user via the device (paragraphs 0002, 0022 user query received);
generating, using an encoder configured to encode a query embedding in a vector space from the text query as a real-valued vector representation of semantic meaning of the text query such that words or sentences from the query that are similar in sematic meaning are closer to each other in the vector space (paragraph 0005 convert query into vector embedding, relevancy of text sample to query, paragraph 0022 relevancy measured with L2-distance metric, paragraph 0034 using an encoder with the vector embeddings);
adding the documents to a reverse index by breaking down unstructured text of the documents into individual sentences while preserving semantic relationships between words and sentences using customized metadata (paragraph 0051 segment vectors in index);
extracting titles from the documents or from metadata associated with the documents or generating titles from contents of the individual sentences, extracting entity tags from the documents using a custom analyzer that isolates entity-specific information such that alphanumeric labels with special characters are not inadvertently separated, parsed, or removed during preprocessing, and generating title embeddings for the titles and sentence embeddings for the individual sentences (paragraphs 0023-0024 generative labeling – create labeled data for embedding model for semantic pairs of content segments, paragraph 0033 content segments of technical specifications (documents) as sentences, paragraph 0034 sentence embedding model for example sentence-BERT);
storing the titles, the title embeddings, the individual sentences, the sentence embeddings, and the entity tags in the reverse index (paragraphs 0034, 0039 storing vector representations (embeddings) sentences, etc.);
scoring entries in the reverse index using a hybrid scoring function, wherein the hybrid scoring function is used to generate scores, each generated score based on five individual scores that comprise keyword matching between (i) the text query and the titles, (ii) the text query and the individual sentences, and (iii) the line of business associated with the user and the entity tags, and cosine similarity between (i) the query embedding and the title embeddings and (ii) the query embedding and the sentence embeddings, and wherein the five individual scores of the hybrid scoring function are computed in parallel using an index structure (paragraphs 0005, 0023 scores of text, sentences similarity with query/question (paragraph 0024), paragraphs 0022, 0041, 0067 cosine similarity for query vs. text in embeddings);
ranking the generated scores for the individual sentences in the document database (paragraph 0022 segments in database ranked per similarity scores); and
displaying an individual sentence associated with a top score, from among the ranked scores, to the user (paragraph 0036 content segment returned as top-K match with high similarity score).
Kotaru does not teach:
adding the documents to a reverse index by breaking down unstructured text of the documents into individual sentences while preserving semantic relationships between words and sentences using customized metadata; and
storing the titles, the title embeddings, the individual sentences, the sentence embeddings, and the entity tags in the reverse index ;
Pearman teaches these things:
adding the documents to a reverse index by breaking down unstructured text of the documents into individual sentences while preserving semantic relationships between words and sentences using customized metadata (paragraph 0031 documents stored in an inverted index which can be a reverse index per specification paragraph 0092)); and
storing the titles, the title embeddings, the individual sentences, the sentence embeddings, and the entity tags in the reverse index (paragraph 0031 content (words, terms) from documents in the inverted index);
It would have been obvious to have combined this function or storing documents and content in a reverse index in Pearman with the techniques for searching documents in Kotaru since both references as from the same assignee Microsoft and using a reverse index would allow for faster searching, making the combination more user-friendly.
With respect to claim 10, Kotaru teaches at least one processor device (paragraphs 0088, 0109, figure 6 processing unit 602), a database for documents (paragraphs 0022, 0032), at least one computer-readable memory device (paragraph88 figure 6 memory 604), and software components (paragraph 0089 figure 6 program modules 606).
With respect top claim 15, all the limitations in claim 10 are addressed by Kotaru and Pearman above. Kotaru also teaches wherein the entity line of business information and the entity tags comprise one or more of: product codes or names, product categories, product types, error codes, lines of business, model numbers, serial numbers, and entity brand names (paragraph 0005 technical specifications (documents) with unique terminologies and acronyms (categories, types), paragraphs 0020, 0032 lines of business such as for an engineer’s role).
Claim 6 is rejected under 35 U.S.C. 103 as being unpatentable over Kotaru and Pearman in further view of Byron et al (US 20170220566), hereafter Byron.
With respect to claim 6, all the limitations in claim 1 are addressed by Kotaru and Pearman above. The combination of Kotaru and Pearman does not teach wherein the user line of business and the entity tags comprise one or more of: product codes or names, product categories, product types, error codes, lines of business, model numbers, serial numbers, and entity brand names. Byron teaches this with a user’s line of business includes information about a user’s role in and organization, and the user’s work history (paragraph 0017). It would have been obvious to have combined this obtaining of additional information in Byron with the techniques for searching for documents in Kotaru and Pearman to provide more information to search which would enhance the relevance of results for a user’s query.
Responses to Applicant’s Remarks
Regarding objection to claim 1 for antecedent basis for the text query in the third limitation, in view of Applicant’s amendments reciting “the line of business associated with the user based on a product or a text query received from a device,” this objection is withdrawn. Regarding rejections to claims 1-19 under 35 U.S.C. 101 for reciting mental processes without significantly more, Applicant’s arguments have been considered but are not persuasive. On pages 14-15 of his Remarks Applicant asserts using an encoder to generate an embedding and adding documents to a reverse index by breaking down unstructured text into individual sentences cannot be performed in the human mind. Examiner disagrees and notes MPEP 2106.04(a)(2)(III) states "The courts do not distinguish between mental processes that are performed entirely in the human mind and mental processes that require a human to use a physical aid (e.g., pen and paper or a slide rule) to perform the claim limitation," and "Nor do the courts distinguish between claims that recite mental processes performed by humans and claims that recite mental processes performed on a computer." Examiner found support in specification paragraph 0093 for an encoder (“embeddings may be created by any present or future Large Language Model (LLM) or encoder”). The limitation recites generating an embedding “using an encoder” which is merely applying it, and is not significantly more than an established method of such software and is not more than an abstract idea per the Recentive Analytics case as shown in the rejection above. Adding documents to a reverse index is recited without details showing how the invention adds the documents and “breaking down unstructured text into individual sentences while preserving relationships” involves evaluation of said sentences and is itself a mental process. Computing scores is also recited broadly (“scoring entries … using a hybrid scoring function”) and without details showing how the invention computes the scores, and a BRI of computing includes simply calculations on paper. Thus Examiner believes these limitations are still mental process steps as shown above.
On pages 16-17 of his Remarks Applicant asserts the claims “recite a particular technological solution to technological problems identified in the Specification: sentence-level reverse indexing with metadata-preserved relationships, title generation when a title is unavailable, custom preprocessing that preserves entity-specific special-character labels, and parallel computation of composite-score components using an index structure.” Examiner disagrees as the amended limitations are still recited broadly and without details from the invention showing how the invention accomplishes the cited improvements. For example, in addition to the limitations discussed above, the claims recite detecting/determining a line of business, extracting titles, entity tags, generating title embeddings, ranking generated scores without details showing how the invention accomplishes the detecting, extracting, generating, and ranking of scores, which may show how the invention improves upon the cited technological problems. Examiner notes specification paragraphs 0066-0067 discusses optimizations in determining shards and nodes as part of the strategy for improvement in reducing latency in searches, which is not claimed. Also on pages 16-17 Applicant discusses Step 2A Prong 2 and Examiner notes the additional element recited are each routine and conventional and do not amount to significantly more than the recited mental processes. For example, receiving documents and a text query are irrelevant to generating embeddings of the query or breaking down the documents into individual sentences, and displaying a sentence with a score is irrelevant to computing and ranking the score. Examiner also notes MPE 2106.05(d)(I) states “The question of whether a particular claimed invention is novel or obvious is ‘fully apart’ from the question of whether it is eligible.”
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Inquiry
Any inquiry concerning this communication or earlier communications from the examiner should be directed to BRUCE M MOSER whose telephone number is (571)270-1718. The examiner can normally be reached M-F 9a-5p.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Boris Gorney can be reached at 571 270-5626. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/BRUCE M MOSER/Primary Examiner, Art Unit 2154 6/18/26
Prosecution Insights