DETAILED ACTION
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 .
This Office action is in response to the amendments, arguments and remarks, filed on 1/6/2026, in which claim(s) 1-15 and 17-21 is/are presented for further examination.
Claim(s) 1, 3, 17, 19 and 20 has/have been amended.
Claim(s) 16 has/have been cancelled.
Claim(s) 21 has/have been added.
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant’s submission filed on 1/6/2026 has been entered.
Response to Amendment
Applicant’s amendment(s) to claim(s) 1, 3, 17, 19 and 20 has/have been accepted.
Applicant’s addition(s) of claim(s) 21 has/have been accepted.
Note: The examiner requests that applicant cite where in the specification there is support for applicant’s amendment(s)/addition(s). It will quicken the prosecution if the examiner does not have to search the entire specification to ensure that applicant has not introduced new matter.
Response to Arguments
Applicant’s arguments with respect to claim(s) 1-15 and 17-21, filed on 1/6/2026, have been fully considered but they are not persuasive.
Applicant’s arguments with respect to the rejection(s) of claim(s) 1-15 and 8-21, under 35 U.S.C. 103, see the bottom of page 11 to the top of page 16 of applicant’s remarks, filed on 1/6/2026, have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Claim Rejections - 35 USC § 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, 12-15 and 17-21 is/are rejected under 35 U.S.C. 103 as being unpatentable over Tiwari et al., US 2021/0133251 A1 (hereinafter “Tiwari”) in view of Neubeck, US 2024/0045846 A1 (hereinafter “Neubeck”) in further view of Sharma et al., US 2021/0406290 A1 (hereinafter “Sharma”).
Claims 1 and 17
Tiwari discloses a method performed within a database system (Tiwari, [0038], see knowledge base 410) comprising:
accessing a plurality of documents (Tiwari, [0025], see the knowledge base corpus 106 includes data in any number of formats, such as CSV files, JSON files, PDF documents, html pages, and the like. In some embodiments, the knowledge base corpus 106 is accessed from websites, databases, and any other data sources, where the data/documents must be first accessed to load/incorporate them into the knowledge base corpus);
for each document of the plurality of documents:
converting data within said each document to plaintext (Tiwari, [0025], see the knowledge base corpus 106 includes data in any number of formats, such as CSV files, JSON files, PDF documents, html pages, and the like. In some embodiments, the knowledge base corpus 106 is accessed from websites, databases, and any other data sources; and Tiwari, [0027], see the knowledge base corpus 106 is accessed from multiple sources. These data sources may be normalized into a common format, such as CSV [i.e., where comma separated values (CSV) is a plain text format] or JSON and mapped to certain fields in an index, … . An example type of document may expect title, description, tags, category, and subcategory fields. If data cannot be mapped to an existing field, then a new type of document index is created);
generating a plurality of chunks based on the plaintext (Tiwari, [0046], see the system uses noun chunks that are automatically extracted from the article or standard system entities (such as cities, colors, location entities, and the like)),
generating, by an embedding model, a plurality of vectors based on the plurality of chunks (Tiwari, [0038], see a vector index 412 includes sentences in the knowledge base 410 embedded into vectors. An index is created for computing similarity, which results in the vector index 412. In some embodiments, sentence embedding is used to create vectors for the document (e.g., title, sentences, and utterances) that are stored in the vector index 412; and Tiwari, [0078], see, given a document, the systems and methods create sentence embeddings for title, each sentence of description, and each of the tags/utterances using a sentence embedding technique. These embeddings are indexed separately (one for each field) using a nearest neighbor technique);
storing the plurality of vectors (See above, see the vector index 412)
storing, a document identifier, that identifies said each document, in association with each vector of the plurality of vectors (Tiwari, [0038], see sentence embedding is used to create vectors for the document (e.g., title, sentences, and utterances) that are stored in the vector index 412, where there must be some identifier/link to associate the created vectors to the corresponding document [i.e., “document identifier”]);
wherein the method is performed by one or more computing devices (Tiwari, [0017], see one or more processors).
Tiwari does not appear to explicitly disclose generating a vector table and a text table as part of a hybrid index that is represented to users as a single interface to access data that is stored in vector form and in textual form;
wherein each chunk in the plurality of chunks corresponds to a different subset of the plaintext;
storing the plurality of vectors in the vector table that comprises a plurality of rows, each row storing a different vector of the plurality of vectors, wherein each vector represents a semantic meaning of a different chunk of the plurality of chunks;
storing, in the vector table, a document identifier;
generating a plurality of tokens based on the plaintext;
storing the plurality of tokens in the text table;
storing, in the text table, the document identifier in association with each token of the plurality of tokens.
Neubeck discloses generating a vector table and a text table (Neubeck, [0024], see an inverted index [i.e., stored in the “text table”, see below] maps a token or term into a position in a document of a repository. In this embodiment, the token represents the content of a document. A token may be a whole word, number, or a sequence of characters; and Neubeck, [0062], see, for each domain 414, 416, there is a global starting positions table 418, 424, a bitrank vector 420 [i.e., “vector table”], 426, and an inverted index table 422, 428 [i.e., “text table”]. The global starting positions table 418, 424 includes the value of the starting position of each global positions block. The bitrank vector 420, 426, is used to perform the hybrid mapping. The inverted index table 422, 428 contains the posting lists for each index);
storing the plurality of vectors in the vector table that comprises a plurality of rows (See Neubeck, [0062] for “vector table” above; Neubeck, [0035], see the bitrank vector 120 is used to map a posting to its respective document and local positions within the document; and Neubeck, Fig. 1, see global starting positions table 114 and inverted index table 116 showing plurality of rows);
storing, in the vector table, a document identifier (See Neubeck, [0062] for “vector table” above; and Neubeck, [0023], see the inverted indexing system uses a global positions block and a bitrank vector to map the single-value posting to a document identifier and the local positions within the corresponding document);
generating a plurality of tokens based on the plaintext (See below);
storing the plurality of tokens in the text table (Neubeck, [0024], see an inverted index [i.e., stored in the “text table”, see Neubeck, [0062] above] maps a token or term into a position in a document of a repository. In this embodiment, the token represents the content of a document. A token may be a whole word, number, or a sequence of characters);
storing, in the text table, the document identifier in association with each token of the plurality of tokens (Neubeck, [0024], see an inverted index [i.e., stored in the “text table”, see Neubeck, [0062] above] maps a token or term into a position in a document of a repository. In this embodiment, the token represents the content of a document. A token may be a whole word, number, or a sequence of characters; and Neubeck, [0026], see the inverted index maps a token, such as a repository property, into a document identifier and a position within a list of repositories associated with the document identifier).
Tiwari and Neubeck are analogous art because they are from the same field of endeavor such as indexing and searching documents.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention, having the teachings of Tiwari and Neubeck before him/her, to modify the indexing of Tiwari to include the tokens of Neubeck because it would allow quick document retrieval.
The suggestion/motivation for doing so would have been to have a hybrid positional posting list to use a single value as a posting or index to represent a document posting and a positional posting, which reduces the size of the posting list, see Neubeck, [0022].
Therefore, it would have been obvious to combine Neubeck with Tiwari to obtain the invention as specified in the instant claim(s).
The combination of Tiwari and Neubeck does not appear to explicitly disclose generating a hybrid index that is represented to users as a single interface to access data that is stored in vector form and in textual form;
wherein each chunk in the plurality of chunks corresponds to a different subset of the plaintext;
storing a different vector of the plurality of vectors, wherein each vector represents a semantic meaning of a different chunk of the plurality of chunks.
Sharma discloses generating a hybrid index that is represented to users as a single interface to access data that is stored in vector form and in textual form (Sharma, [0055], see the textual search logic 50 compares the text of an argument paragraph to the text from index items in the corpus index 38c [i.e., corresponds to the “textual form index”] associated with documents in the document database 38 b. In some examples, the textual search logic 50 performs the textual search based on a portion of the text of the argument paragraph including only certain words or phrases therein; Sharma, [0057], see the semantic search logic 52 may determine a vector representation or embedding of an argument paragraph, which may represent a semantic meaning of the argument paragraph. In some examples, the semantic search logic 52 may use Bidirectional Encoder Representations from Transformers (BERT) to determine a vector representation of an argument paragraph. In other examples, other natural language processing tools may be used to determine a vector representation of an argument paragraph. A vector representation of index items of the corpus index 38c [i.e., corresponds to the ”vector form index”]; and Sharma, [0060], see, after the textual search logic 50 performs a textual search and the semantic search logic 52 performs a semantic search, two result sets are available. One result set is based on the textual search and one result set is based on the semantic search. As discussed above, these two searches may yield different results. Thus, combining the two search results may yield more useful results [i.e., corresponds the “single user interface to access data”] than either search result individually);
wherein each chunk in the plurality of chunks corresponds to a different subset of the plaintext (Sharma, [0053], see the textual search logic 50 performs a search for legal references relevant to an argument paragraph [i.e., corresponds to a “chunk”] based on the text of the argument paragraph. Specifically, the textual search logic 50 performs keyword based searching. In particular, the textual search logic 50 compares the text of an argument paragraph against the text in the corpus index 38c.);
storing a different vector of the plurality of vectors, wherein each vector represents a semantic meaning of a different chunk of the plurality of chunks (Sharma, [0057], see the semantic search logic 52 may determine a vector representation or embedding of an argument paragraph, which may represent a semantic meaning of the argument paragraph. In some examples, the semantic search logic 52 may use Bidirectional Encoder Representations from Transformers (BERT) to determine a vector representation of an argument paragraph. In other examples, other natural language processing tools may be used to determine a vector representation of an argument paragraph. A vector representation of index items of the corpus index 38c [i.e., corresponds to the ”vector form index”]).
Tiwari, Neubeck and Sharma are analogous art because they are from the same field of endeavor such as indexing and searching documents.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention, having the teachings of Tiwari, Neubeck and Sharma before him/her, to modify the token indexing of the combination of Tiwari and Neubeck to include the semantic searching of Sharma because it would obtain more precise results.
The suggestion/motivation for doing so would have been to automatically analyze and perform searches, see Sharma, [0016].
Therefore, it would have been obvious to combine Sharma with the combination of Tiwari and Neubeck to obtain the invention as specified in the instant claim(s).
Claim(s) 17 recite(s) similar limitations to claim 1 and is/are rejected under the same rationale.
With respect to claim 17, Tiwari discloses one or more non-transitory storage media storing instructions (Tiwari, [0018], see computer storage media).
Claims 2 and 18
With respect to claims 2 and 18, the combination of Tiwari, Neubeck and Sharma discloses wherein:
the plurality of documents comprise a first document of a first type and a second document of a second type that is different than the first type (See below);
converting the data with said each document comprises:
for the first document of the plurality of documents, performing a first conversion operation on the first document based on the first document being of the first type (See below);
for the second document of the plurality of documents, performing a second conversion operation on the second document based on the second document being of the second type (Tiwari, [0025], see the knowledge base corpus 106 includes data in any number of formats, such as CSV files, JSON files, PDF documents, html pages, and the like. In some embodiments, the knowledge base corpus 106 is accessed from websites, databases, and any other data sources; and Tiwari, [0027], see the knowledge base corpus 106 is accessed from multiple sources. These data sources may be normalized into a common format, such as CSV [i.e., where comma separated values (CSV) is a plain text format] or JSON and mapped to certain fields in an index, … . An example type of document may expect title, description, tags, category, and subcategory fields. If data cannot be mapped to an existing field, then a new type of document index is created, where there is different mappings into the indices based on whether the file being a CSV file, JSON file, PDF document, html page, etc. and the corresponding mapping is used to convert the respective file).
Claims 3 and 19
With respect to claims 3 and 19, the combination of Tiwari, Neubeck and Sharma discloses further comprising:
receiving a database statement that instructs a database system to create the hybrid index (Tiwari, [0024], see creating an index of information for answering questions; Neubeck, [0004], see an information retrieval system uses an inverted indexing system composed of hybrid positional posting lists. A hybrid positional posting list includes a number of postings, where each posting contains a single numeric value to represent both a document and the location of a term in the document; and Neubeck, Fig. 5, see “Start” and creating the inverted index system with hybrid positional posting lists);
wherein generating the vector table and the text table is performed in response to receiving the database statement (Neubeck, Fig. 5, see “Start” and creating the inverted index system with hybrid positional posting lists).
Claims 4 and 20
With respect to claims 4 and 20, the combination of Tiwari, Neubeck and Sharma discloses further comprising:
generating a vector index on the vector table in response to receiving the database statement (Tiwari, [0038], see a vector index 412 includes sentences in the knowledge base 410 embedded into vectors. An index is created for computing similarity, which results in the vector index 412; and Neubeck, Fig. 5, see “Start” and creating the inverted index system with hybrid positional posting lists);
generating a text index on the text table in response to receiving the database statement (Neubeck, Fig. 5, see “Start” and creating the inverted index system with hybrid positional posting lists).
Claim 5
With respect to claim 5, the combination of Tiwari, Neubeck and Sharma discloses wherein the database statement indicates the embedding model (Tiwari, [0072], see the filtering step uses Universal Sentence Encoder embeddings to compute similarity between the original sentence and the generated sentence, and it removes very similar or very dissimilar paraphrases from the pool. The systems and methods run two algorithms sequentially for de-duplicating the pool of paraphrases obtained after filtering).
Claim 6
With respect to claim 6, the combination of Tiwari, Neubeck and Sharma discloses wherein the database statement indicates a type of vector index, the method further comprising:
generating the type of vector index in response to receiving the database statement (Tiwari, [0038], see a vector index 412 includes sentences in the knowledge base 410 embedded into vectors. An index is created for computing similarity, which results in the vector index 412; and Neubeck, Fig. 5, see “Start” and creating the inverted index system with hybrid positional posting lists).
Claim 7
With respect to claim 7, the combination of Tiwari, Neubeck and Sharma discloses wherein the database statement indicates a technique for chunking the plaintext, wherein generating the plurality of chunks is performed using the technique (Tiwari, [0046], see the system uses noun chunks that are automatically extracted from the article or standard system entities (such as cities, colors, location entities, and the like)).
Claim 8
With respect to claim 8, the combination of Tiwari, Neubeck and Sharma discloses wherein the database statement indicates a distance operation for computing a distance between two vectors (Tiwari, [0072], see the filtering step uses Universal Sentence Encoder embeddings to compute similarity [i.e., “distance”] between the original sentence and the generated sentence, and it removes very similar or very dissimilar paraphrases from the pool. The systems and methods run two algorithms sequentially for de-duplicating the pool of paraphrases obtained after filtering).
Claim 9
With respect to claim 9, the combination of Tiwari, Neubeck and Sharma further comprising:
receiving a hybrid query that indicates one or more search terms (Neubeck, [0054], see the hybrid positional posting list is used to search for documents using postings from different domains. Each domain uses distinct tokens. In the content domain, the inverted index is a token that represents the content of a document and in the repository domain, the inverted index is a repository property, which the query for the system would be a hybrid query; and Neubeck, [0055], see a query may contain search terms from different domains, such as a search term from the content domain and a search term from a repository domain. The inverted mapping technique is able to search from one domain and to another domain using the mapping components in order to more readily find documents matching the postings that intersect both domains);
generating one or more query vectors based on the one or more search terms (See Neubeck, [0054] and [0055] above for the query; and Neubeck, [0062], see, for each domain 414, 416, there is a global starting positions table 418, 424, a bitrank vector 420, 426, and an inverted index table 422, 428. The global starting positions table 418, 424 includes the value of the starting position of each global positions block. The bitrank vector 420, 426, is used to perform the hybrid mapping. The inverted index table 422, 428 contains the posting lists for each index, which means the query need to be converted into bitrank vectors in order to search the inverted index table for the corresponding bitrank vector);
identifying, based on the one or more query vectors, one or more document identifiers from the vector table (Neubeck, [0056], see the cross-domain mapping 300 for the search operation “content: abc OR repo: OwnerID7” 302. The cross-domain mapping includes searching for the term ‘abc’ in the content domain 306 and searching for the term ‘OwnerID7’ in the repository domain 308. The search may use the global postings for the term ‘abc’ 312 in the inverted index table of the content domain 310 to map, using the hybrid positional posting mapping 314, to a document containing the term, DocID1, DocID2, . . . DocID12 316. Each document is mapped into its associated repositories 318 using a mapping 320 that includes the global starting positions table of the repository domain 322 and the bitrank vector of the repository domain 324, see using the bitrank vector to return documents DocID1, DocID2, . . . DocID12 316).
Claim 10
With respect to claim 10, the combination of Tiwari, Neubeck and Sharma wherein the one or more search terms are one or more first search terms, the method further comprising:
in response to receiving the hybrid query, generating a text query that includes one or more second search terms, wherein the one or more second search terms are either (i) one or more of the one or more first search terms or (ii) different than the one or more first search terms (See below);
executing the text query, wherein executing the text query comprises identifying, based on the one or more second search terms, one or more second document identifiers from the text table (Neubeck, [0056], see the cross-domain mapping 300 for the search operation “content: abc OR repo: OwnerID7” 302. The cross-domain mapping includes searching for the term ‘abc’ [i.e., “one or more of the one or more first search terms”] in the content domain 306 and searching for the term ‘OwnerID7’ [i.e., “different than the one or more first search terms”] in the repository domain 308. The search may use the global postings for the term ‘abc’ 312 in the inverted index table of the content domain 310 to map, using the hybrid positional posting mapping 314, to a document containing the term, DocID1, DocID2, . . . DocID12 316. Each document is mapped into its associated repositories 318 using a mapping 320 that includes the global starting positions table of the repository domain 322 and the bitrank vector of the repository domain 324, see using the bitrank vector to return documents DocID1, DocID2, . . . DocID12 316).
Claim 12
With respect to claim 12, the combination of Tiwari, Neubeck and Sharma discloses further comprising:
using the one or more query vectors to identify a plurality of chunks in the vector table (See below);
determining that two or more chunks in the plurality of chunks belong to a particular document of the plurality of documents (Neubeck, [0024], see an inverted index [i.e., stored in the “text table”, see below] maps a token or term into a position in a document of a repository. In this embodiment, the token represents the content of a document. A token may be a whole word, number, or a sequence of characters; and Neubeck, [0062], see, for each domain 414, 416, there is a global starting positions table 418, 424, a bitrank vector 420 [i.e., “vector table”], 426, and an inverted index table 422, 428 [i.e., “text table”]. The global starting positions table 418, 424 includes the value of the starting position of each global positions block. The bitrank vector 420, 426, is used to perform the hybrid mapping. The inverted index table 422, 428 contains the posting lists for each index; and Neubeck, [0056], see the cross-domain mapping 300 for the search operation “content: abc OR repo: OwnerID7” 302. The cross-domain mapping includes searching for the term ‘abc’ [i.e., “one or more of the one or more first search terms”] in the content domain 306 and searching for the term ‘OwnerID7’ [i.e., “different than the one or more first search terms”] in the repository domain 308. The search may use the global postings for the term ‘abc’ 312 in the inverted index table of the content domain 310 to map, using the hybrid positional posting mapping 314, to a document containing the term, DocID1, DocID2, . . . DocID12 316. Each document is mapped into its associated repositories 318 using a mapping 320 that includes the global starting positions table of the repository domain 322 and the bitrank vector of the repository domain 324, see using the bitrank vector to return documents DocID1, DocID2, . . . DocID12 316);
for a first chunk of the two or more chunks:
identifying a first plurality of chunk scores of chunks that surround the first chunk;
generating, for the first chunk, a first adjustment score based on the first plurality of chunk scores and a first chunk score of the first chunk (Tiwari, [0043], see the described systems and methods provide a transformed query that removes stop words and intent keywords from the user message. …, the systems and methods described herein use various approaches to match the indexed data with the query. When scoring, the systems and methods may specify a scoring function for each document-query combination. …, this scoring may be performed in vector space [i.e., “first adjustment score”]. After scoring the documents, the systems and methods may rank the results of the scoring; and Tiwari, [0045], see the systems and methods use fields such as the title and description of the article to retrieve the fields. … , metadata is used to help the systems and methods determine which article to retrieve for a particular type of user query, thereby improving accuracy of retrieving the correct document [i.e., “second adjustment score”]. … During the scoring phase, each tag is treated uniquely. With vector scoring, the systems and methods select the best matched tag for comparing two articles, where the above process is run for each data chunk);
for a second chunk of the two or more chunks:
identifying a second plurality of chunk scores of chunks that surround the second chunk (See below);
generating, for the second chunk, a second adjustment score based on the second plurality of chunk scores and a second chunk score of the second chunk (Tiwari, [0043], see the described systems and methods provide a transformed query that removes stop words and intent keywords from the user message. …, the systems and methods described herein use various approaches to match the indexed data with the query. When scoring, the systems and methods may specify a scoring function for each document-query combination. …, this scoring may be performed in vector space [i.e., “first adjustment score”]. After scoring the documents, the systems and methods may rank the results of the scoring; and Tiwari, [0045], see the systems and methods use fields such as the title and description of the article to retrieve the fields. … , metadata is used to help the systems and methods determine which article to retrieve for a particular type of user query, thereby improving accuracy of retrieving the correct document [i.e., “second adjustment score”]. … During the scoring phase, each tag is treated uniquely. With vector scoring, the systems and methods select the best matched tag for comparing two articles, where the above process is run for each data chunk);
determining a score for the particular document based on the first adjustment score and the second adjustment score (See above the combination of the scoring function for each document-query combination and the scoring using metadata to improve the accuracy of retrieving the correct document).
Claim 13
With respect to claim 13, the combination of Tiwari, Neubeck and Sharma discloses further comprising:
using the one or more query vectors to identify a set of chunks in the vector table (See below);
determining that a plurality of chunks in the set of chunks belong to a particular document of the plurality of documents (See below);
identifying a window size that is less than the number of chunks in the plurality of chunks (See below);
identifying a plurality of subsets of the plurality of chunks (Neubeck, [0054], see the hybrid positional posting list is used to search for documents using postings from different domains. Each domain uses distinct tokens. In the content domain, the inverted index is a token that represents the content of a document and in the repository domain, the inverted index is a repository property, which the query for the system would be a hybrid query; Neubeck, [0055], see a query may contain search terms from different domains, such as a search term from the content domain and a search term from a repository domain. The inverted mapping technique is able to search from one domain and to another domain using the mapping components in order to more readily find documents matching the postings that intersect both domains; Neubeck, [0062], see, for each domain 414, 416, there is a global starting positions table 418, 424, a bitrank vector 420, 426, and an inverted index table 422, 428. The global starting positions table 418, 424 includes the value of the starting position of each global positions block. The bitrank vector 420, 426, is used to perform the hybrid mapping. The inverted index table 422, 428 contains the posting lists for each index, which means the query need to be converted into bitrank vectors in order to search the inverted index table for the corresponding bitrank vector; and Neubeck, Fig. 8);
for each subset in the plurality of subsets:
generating an average score of chunks scores of chunks in said each subset (See below);
adding the average score to a set of average scores (Tiwari, [0039], see optimizes and tunes the weights for various feature scores (e.g., text score, vector score, title similarity, utterance similarity, etc.) to combine those score features);
identifying the maximum average score in the set of average scores (Tiwari, [0042], see, if at least one article is determined to be above the confidence threshold level, then the top article (e.g., the highest ranked article) is returned 524 to the user);
associating the maximum average score with the particular document (Tiwari, [0042], see, if at least one article is determined to be above the confidence threshold level, then the top article (e.g., the highest ranked article) [i.e., associated with the particular document/article] is returned 524 to the user).
Claims 14 and 21
With respect to claims 14 and 21, the combination of Tiwari, Neubeck and Sharma discloses further comprising:
receiving a hybrid query that indicates one or more search terms (Sharma, [0051], see identifies one or more argument paragraphs [i.e., corresponds to the “one or more search terms”], a search may be performed for the identified argument paragraphs);
in response to receiving the hybrid query, generating (1) a text sub-query that targets the text table (See Sharma, [0055] below) and (2) a vector sub-query that targets the vector table (See Sharma, [0057] below);
executing the text sub-query to generate a first set of results (Sharma, [0055], see the textual search logic 50 compares the text of an argument paragraph to the text from index items in the corpus index 38c [i.e., corresponds to the “textual form index”] associated with documents in the document database 38 b. In some examples, the textual search logic 50 performs the textual search based on a portion of the text of the argument paragraph including only certain words or phrases therein);
executing the vector sub-query to generate a second set of results (Sharma, [0057], see the semantic search logic 52 may determine a vector representation or embedding of an argument paragraph, which may represent a semantic meaning of the argument paragraph. In some examples, the semantic search logic 52 may use Bidirectional Encoder Representations from Transformers (BERT) to determine a vector representation of an argument paragraph. In other examples, other natural language processing tools may be used to determine a vector representation of an argument paragraph. A vector representation of index items of the corpus index 38c [i.e., corresponds to the ”vector form index”]);
combining the first set of results with the second set of results to generate a final set of results (Sharma, [0060], see, after the textual search logic 50 performs a textual search and the semantic search logic 52 performs a semantic search, two result sets are available. One result set is based on the textual search and one result set is based on the semantic search. As discussed above, these two searches may yield different results. Thus, combining the two search results may yield more useful results [i.e., corresponds the “single user interface to access data”] than either search result individually).
Claim 15
With respect to claim 15, the combination of Tiwari, Neubeck and Sharma discloses wherein the first set of results is a first set of document identifiers, wherein the second set of results is a second set of document identifiers, wherein the final set of results is (i) a union of the first set of document identifiers and the second set of document identifiers or (ii) an intersection of the first set of document identifiers and the second set of document identifiers (Neubeck, Fig. 8, see searching the hybrid positional posting list for each token [i.e., the text-subquery and vector-subquery] and the combined results returned).
Claim(s) 11 is/are rejected under 35 U.S.C. 103 as being unpatentable over Tiwari in view of Neubeck in further view of Sharma in further view of Garcia Calatrava et al., US 2023/0418800 A1 (hereinafter “Garcia”).
Claim 11 incorporates all of the limitations above.
The combination of Tiwari, Neubeck and Sharma discloses wherein the hybrid query is a first hybrid query (Sharma, [0051], see identifies one or more argument paragraphs [i.e., corresponds to the “one or more search terms”], a search may be performed for the identified argument paragraphs),
or a low SQL query that includes both and non-SQL elements (Tiwari, [0078], see the systems and methods convert the query to a vector using a sentence embedding technique. The systems and methods then search for the query vector in each of the field vectors. This is a search in the vector space [i.e., using BRI, “non-SQL elements”] and uses approximate nearest neighbor to optimize for performance).
The combination of Tiwari, Neubeck and Sharma does not appear to explicitly disclose wherein the first hybrid query is a SQL query, the method further comprising:
receiving a second hybrid query that indicates the hybrid index, wherein the second hybrid query is a NOSQL query or a low SQL query that includes SQL elements.
Garcia discloses wherein the first hybrid query is a SQL query (Garcia, [0015], see users using InfluxQL, a SQL-like query language), the method further comprising:
receiving a second hybrid query that indicates the hybrid index (Garcia, [0092], see the bridge data model is intended to optimize hybrid queries), wherein the second hybrid query is a NOSQL query (Garcia, [0108] and [0115], see the method has been evaluated on MongoDB 5.0 CE, the most popular NoSQL database) or a low SQL query that includes SQL elements (Garcia, [0130], see, in SQL terms, all querying types could consist only in three different clauses: SELECT, FROM and WHERE, except from the aggregation querying ones, that could also incorporate a GROUP BY clause).
Tiwari, Neubeck, Sharma and Garcia are analogous art because they are from the same field of endeavor such as indexing and searching documents.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention, having the teachings of Tiwari, Neubeck, Sharma and Garcia before him/her, to modify the token indexing of the combination of Tiwari, Neubeck and Sharma to include the use of NoSQL of Garcia because it would allow native time-series capabilities/searching.
The suggestion/motivation for doing so would have been to tailor the database not only to time series data, but also to the natural data-flow of real-time data (ingestion, storage, retrieval), see Garcia, [0022].
Therefore, it would have been obvious to combine Garcia with the combination of Tiwari, Neubeck and Sharma to obtain the invention as specified in the instant claim(s).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
– Dexter et al., 2009/0216736 for displaying document chunks in response to a search request;
– Turkkan et al., 12299397 for determining semantic similarity of texts based on sub-sections thereof;
– Dexter et al., 8751484 for identifying chunks within multiple documents;
– Li et al., 12417249 for smart find for in-application searching;
– Madisetti et al., 12405978 for optimizing use of retrieval augmented generation pipelines in generative AI applications; and
– Jiang et al., WO 2018040503 for 8793265 for obtaining search results.
Point of Contact
Any inquiry concerning this communication or earlier communications from the examiner should be directed to HUBERT G CHEUNG whose telephone number is (571) 270-1396. The examiner can normally be reached M-R 8:00A-5:00P EST; alt. F 8:00A-4:00P EST.
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, Neveen Abel-Jalil can be reached at (571) 270-0474. 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.
Examiner: Hubert Cheung
/Hubert Cheung/Assistant Examiner, Art Unit 2152Date: March 23, 2026
/NEVEEN ABEL JALIL/Supervisory Patent Examiner, Art Unit 2152
Prosecution Insights