DETAILED ACTION
Remarks
This Office Action is in response to the application 19/066050 filed on 27 February 2025.
Claims 1-20 have been examined.
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 .
Double Patenting
The nonstatutory double patenting rejection is based on a judicially created doctrine grounded in public policy (a policy reflected in the statute) so as to prevent the unjustified or improper timewise extension of the “right to exclude” granted by a patent and to prevent possible harassment by multiple assignees. A nonstatutory double patenting rejection is appropriate where the conflicting claims are not identical, but at least one examined application claim is not patentably distinct from the reference claim(s) because the examined application claim is either anticipated by, or would have been obvious over, the reference claim(s). See, e.g., In re Berg, 140 F.3d 1428, 46 USPQ2d 1226 (Fed. Cir. 1998); In re Goodman, 11 F.3d 1046, 29 USPQ2d 2010 (Fed. Cir. 1993); In re Longi, 759 F.2d 887, 225 USPQ 645 (Fed. Cir. 1985); In re Van Ornum, 686 F.2d 937, 214 USPQ 761 (CCPA 1982); In re Vogel, 422 F.2d 438, 164 USPQ 619 (CCPA 1970); In re Thorington, 418 F.2d 528, 163 USPQ 644 (CCPA 1969).
A timely filed terminal disclaimer in compliance with 37 CFR 1.321(c) or 1.321(d) may be used to overcome an actual or provisional rejection based on nonstatutory double patenting provided the reference application or patent either is shown to be commonly owned with the examined application, or claims an invention made as a result of activities undertaken within the scope of a joint research agreement. See MPEP § 717.02 for applications subject to examination under the first inventor to file provisions of the AIA as explained in MPEP § 2159. See MPEP § 2146 et seq. for applications not subject to examination under the first inventor to file provisions of the AIA . A terminal disclaimer must be signed in compliance with 37 CFR 1.321(b).
The filing of a terminal disclaimer by itself is not a complete reply to a nonstatutory double patenting (NSDP) rejection. A complete reply requires that the terminal disclaimer be accompanied by a reply requesting reconsideration of the prior Office action. Even where the NSDP rejection is provisional the reply must be complete. See MPEP § 804, subsection I.B.1. For a reply to a non-final Office action, see 37 CFR 1.111(a). For a reply to final Office action, see 37 CFR 1.113(c). A request for reconsideration while not provided for in 37 CFR 1.113(c) may be filed after final for consideration. See MPEP §§ 706.07(e) and 714.13.
The USPTO Internet website contains terminal disclaimer forms which may be used. Please visit www.uspto.gov/patent/patents-forms. The actual filing date of the application in which the form is filed determines what form (e.g., PTO/SB/25, PTO/SB/26, PTO/AIA /25, or PTO/AIA /26) should be used. A web-based eTerminal Disclaimer may be filled out completely online using web-screens. An eTerminal Disclaimer that meets all requirements is auto-processed and approved immediately upon submission. For more information about eTerminal Disclaimers, refer to www.uspto.gov/patents/apply/applying-online/eterminal-disclaimer.
Claim 1 is provisionally rejected on the ground of nonstatutory double patenting as being unpatentable over claim 1 of copending U.S. application no. 19/068993 in view of Kotaru et al. (U.S. Patent Application Publication No. 20240330589 A1, hereinafter referred to as Kotaru). Although they are not identically worded, claim 1 of the reference application teaches almost all of the features of examined claim 1. Reference claim 1 performs “augmenting” of a user query based on vector embeddings, which can be considered a type of “extracted additional information.” Examined claim 1 generates a second query based on the first query and using extracted additional information. This corresponds to augmenting the first query with the extracted additional information. Although reference claim 1 fails to mention any “semantic similarity,” “relevance score,” or ranking elements based on these; Kotaru teaches this (see Kotaru para. 0067-0070, and see the claim rejections under 35 U.S.C., detailed below). It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to have modified reference claim 1 to include the teachings of Kotaru because it allows efficient retrieval of the most relevant database data to provide context for a user’s query (Kotaru para. 0067-0070).
This is a provisional nonstatutory double patenting rejection. If the co-pending application(s) are issued as patent(s), the double patenting rejections will be converted from provisional to non-provisional.
Claim 1 is provisionally rejected on the ground of nonstatutory double patenting as being unpatentable over claims 1, 6, and 7 of copending U.S. application no. 19/079190 in view of Kotaru et al. (U.S. Patent Application Publication No. 20240330589 A1, hereinafter referred to as Kotaru). Although they are not identically worded, claims 1, 6, and 7 of the reference application teach almost all of the features of examined claim 1, except for the claimed ranking and question answer system. Kotaru teaches these features (see Kotaru para. 0054, 0067-0070, and Fig. 2; and see the claim rejections under 35 U.S.C., detailed below). It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to have modified reference claims 1, 6, and 7 to include the teachings of Kotaru because it allows efficient retrieval of the most relevant database data to provide context for a user’s query (Kotaru para. 0067-0070).
This is a provisional nonstatutory double patenting rejection. If the co-pending application(s) are issued as patent(s), the double patenting rejections will be converted from provisional to non-provisional.
Claim Rejections - 35 USC § 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 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
As to claims 1, 8, and 15, these claims recite a first user query and a corresponding set of elements comprising one or more documents, one or more knowledge graph triplets, or both. The broadest reasonable interpretation (BRI) of this limitation encompasses a simple case of having a user query and just two corresponding documents. These claims recite “assigning, based at least in part on a respective semantic similarity between the first user query and each element of the set of elements, a respective relevance score to each element of the set of elements” (claim 1 and similar limitations of claims 8 and 15). Given that the BRI of the claims encompasses such a simple case, a human could mentally perform the claimed assigning with the aid of pencil and paper. For example, a human could mentally judge/evaluate the two documents in relation to the query and assign corresponding relevance scores to the documents accordingly. If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind (and/or with a pencil and paper) but for the recitation of generic computer components, then it falls within the “Mental Processes” grouping of abstract ideas. Alternatively, this limitation may be deemed an abstract idea under the “Mathematical Concepts” grouping because the claimed “relevance score” is a numerical quantity and the claimed “assigning” may involve determining the relevance scores for the documents based on mathematical calculation(s).
These claims also recite “ranking the set of elements based at least in part on the respective relevance score assigned to each element of the set of elements” (claim 1 and similar limitations of claims 8 and 15). Given that the BRI of the claims encompasses a simple case, as set forth above, a human could mentally perform the claimed ranking with the aid of pencil and paper. A human can easily mentally rank two documents based on their corresponding relevance scores. Hence, this limitation is also an abstract idea under the “Mental Processes” grouping.
These claims also recite “extracting additional information from each element of the set of elements based at least in part on the first user query” (claim 1 and similar limitations of claims 8 and 15). Given that the BRI of the claims encompasses a simple case, as set forth above, a human could mentally perform the claimed extracting with the aid of pencil and paper. The claims do not specify nor place any limits upon the type of documents, their size, or their length. The claimed “extracting” amounts to no more than a series of mentally-performable judgements/evaluations, i.e. judging/evaluating which information is relevant to the query and hence should be extracted as “additional information.” Hence, this limitation is also an abstract idea under the “Mental Processes” grouping.
These claims also recite “generating a second query based at least in part on the first user query and using the extracted additional information and the ranked set of elements” (claim 1 and similar limitations of claims 8 and 15). Given that the BRI of the claims encompasses a simple case, as set forth above, a human could mentally perform the claimed generating with the aid of pencil and paper. Looking at the first user query, the extracted additional information, and the set of two documents, a human could mentally generated a second query, e.g. one that asks for further details or context. Hence, this limitation is also an abstract idea under the “Mental Processes” grouping. Accordingly, the claim recites an abstract idea.
This judicial exception is not integrated into a practical application. Other than the abstract idea, the claims recite the following:
a) “receiving at a question and answer (QA) system, a first user query” (claim 1 and similar limitations of claims 8 and 15);
b) “obtaining, from one or more databases based at least in part on the first user
query, a set of elements comprising one or more documents, one or more knowledge graph triplets, or both” (claim 1 and similar limitations of claims 8 and 15);
c) “inputting the generated second query into a large language model (LLM)” (claim 1 and similar limitations of claims 8 and 15);
d) “returning an answer generated by the LLM as a response to the first user
query” (claim 1 and similar limitations of claims 8 and 15);
e) “one or more memories storing processor-executable code” (claim 8);
f) “one or more processors coupled with the one or more memories and
individually or collectively operable to execute the code” (claim 8);
g) “A non-transitory computer-readable medium storing code” (claim 15).
Limitations (a) and (b) amount to no more than mere data gathering, which has been deemed by the courts to be insignificant extra-solution activity. See MPEP 2106.05(g). Limitation (c) is recited as a high level of generality and amounts to mere instructions to apply the abstract on a general purpose computer, which cannot provide a practical application. See MPEP 2106.05(f). Limitation (d) amounts to no more than merely outputting a result, which has been deemed by the courts to be insignificant extra-solution activity. See Mayo, 566 U.S. at 79, 101 USPQ2d at 1968; OIP Techs., Inc. v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1092-93 (Fed. Cir. 2015); Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350, 1354-55, 119 USPQ2d 1739, 1742 (Fed. Cir. 2016). See MPEP 2106.05(g). Limitations (e) through (g) are recited at a high level of generality, i.e. as generic computer components performing generic computing functions. Accordingly, these additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. Looking at the additional elements as a whole adds nothing beyond the additional elements considered individually—they still represent insignificant extra-solution activity and/or generic computer implementation. Hence, the claim as a whole, looking at the additional elements individually and in combination, does not integrate the abstract idea into a practical application. The claim is directed to an abstract idea.
The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception. Limitations (a) and (b) amount to no more than mere data gathering, which has been deemed by the courts to be insignificant extra-solution activity. See MPEP 2106.05(g). In addition, the courts have deemed receiving data to be well-understood, routine, and conventional activity, as in the following cases: Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information); OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network); buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014) (computer receives and sends information over a network); Versata Dev. Group, Inc. v. SAP Am., Inc., 793 F.3d 1306, 1334, 115 USPQ2d 1681, 1701 (Fed. Cir. 2015) (storing and retrieving information in memory). See MPEP 2106.05(d)(II). Hence, elements (a) and (b) cannot be deemed an inventive concept. Limitation (c) is recited as a high level of generality and amounts to mere instructions to apply the abstract on a general purpose computer, which cannot be deemed an inventive concept. See MPEP 2106.05(f). Limitation (d) amounts to no more than merely outputting a result, which has been deemed by the courts to be insignificant extra-solution activity. See Mayo, 566 U.S. at 79, 101 USPQ2d at 1968; OIP Techs., Inc. v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1092-93 (Fed. Cir. 2015); Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350, 1354-55, 119 USPQ2d 1739, 1742 (Fed. Cir. 2016). See MPEP 2106.05(g). Furthermore, Applicant’s specification provides few details about the claimed “returning an answer” or its functions (see para. 0096 of Applicant’s published specification). This indicates that this feature is well known in the art. Cf Hybritech Inc. v. Monoclonal Antibodies, Inc., 802 F.2d 1367, 1384 (Fed. Cir. 1986) (explaining that "a patent need not teach, and preferably omits, what is well known in the art"). As a result, the written description adequately supports that additional element (d) is conventional and performs well-understood, routine, and conventional activities. See MPEP § 2106.07(a)(III)(A)1. As discussed above with respect to integration of the abstract idea into a practical application, additional elements (e) through (g) amount to no more than mere field of use limitations and instructions to apply the exception using generic computer components. Mere instructions to apply an exception using conventional computer components and functions cannot provide an inventive concept. Looking at the additional elements as a whole adds nothing beyond the additional elements considered individually—they still represent insignificant extra-solution activity; well-understood, routine, and conventional subject matter; and/or generic computer implementation. Hence, the claim as a whole, looking at the additional elements individually and in combination, does not amount to significantly more than the abstract idea. These claims are not patent eligible.
As to dependent claims 2-4, 9-11, and 16-18, these claims recite a directed acyclic graph (DAG) structure configured to generate the second query based on the first query. The DAG structure includes a first or second set of node depending on whether the QA system is in a smart mode or a thoughtful mode. Beyond this, these claims do not specify nor place any limits upon the DAG structure. Under the BRI, the DAG structure is a simple graph having just a few nodes and edges. Given that the BRI encompasses such a simple case, a human can mentally visualize such a DAG structure and draw it out on a piece of paper. Since nothing in these claims goes beyond what a human could mentally perform with the aid of pencil and paper, these claims are also directed to an abstract idea under the “Mental Processes” grouping, without significantly more.
As to dependent claims 5, 12, and 19, these claims recite “obtaining,” “assigning,” “ranking,” and “extracting” steps that are analogous to those recited in the parent claism (i.e. claims 1, 8, and 15). These “obtaining,” “assigning,” “ranking,” and “extracting” limitations are directed to an abstract idea without significantly more for the same reasons set forth above in the discussion of the parent claims.
As to dependent claims 6, 13, and 20, these claims recite “the one or more know ledge graph tiiplets are obtained from one or more knowledge graphs generated by the LLM from a corpus of documents, and each knowledge graph triplet comprises a subject element, a relationship element, and an object element” (claim 6, and similar limitations of claims 13 and 20). These limitations amount to generally linking the use of the abstract to a particular field of use and/or technological environment, which cannot provide a practical application nor an inventive concept. See MPEP 2106.05(h).
As to dependent claims 7 and 14, these claims recite “wherein inputting the generated second query into the LLM comprises: generating a response template comprising the generated second query, the first user query, the extracted additional infonnation, and the ranked set of elements;, wherein the response template includes formatting elements” (claim 7 and similar limitations of claim 14). Given that the BRI of the claims encompasses a simple case, as set forth above, a human could, with the aid of pencil and paper, mentally perform the claimed generating of a response template. Hence, this limitation is an abstract idea under the “Mental Processes” grouping. These claims also recite: “inputting the response template into the LLM, wherein the LLM generates the answer based on the response template.” This additional element is recited as a high level of generality and amounts to mere instructions to apply the abstract on a general purpose computer, which cannot be deemed an inventive concept. See MPEP 2106.05(f). Looking at the additional elements as a whole adds nothing beyond the additional elements considered individually—they still represent insignificant extra-solution activity; well-understood, routine, and conventional subject matter; and/or generic computer implementation. Hence, the claims as a whole, looking at the additional elements individually and in combination, does not amount to a practical application nor an inventive concept. These claims are not patent eligible.
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 of this title, 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.
Claims 1, 7, 8, 14, and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Somech et al. (U.S. Patent Application Publication No. 20240311563 A1, hereinafter referred to as Somech) in view of Kotaru et al. (U.S. Patent Application Publication No. 20240330589 A1, hereinafter referred to as Kotaru).
As to claim 1, Somech teaches a method, comprising:
receiving at a question and answer (QA) system, a first user query (Somech para. 0003: a user issues a query, such as “when is Red Sea due?”);
obtaining, from one or more databases based at least in part on the first user query, a set of elements comprising one or more documents, one or more knowledge graph triplets, or both (Somech para. 0003: corpus data supplement, e.g. a meeting transcript document);
assigning a respective relevance score to each element of the set of elements (Somech para. 0053: ranking based on relevance score);
ranking the set of elements based at least in part on the respective relevance score assigned to each element of the set of elements (Somech para. 0053: ranking based on relevance score);
extracting additional information from each element of the set of elements based at least in part on the first user query (Somech para. 0003: the corpus data supplement provides contextual data or metadata);
generating a second query based at least in part on the first user query and using the extracted additional information and the ranked set of elements (Somech para. 0003: the contextual data or metadata is input into a large language model (LLM));
inputting the generated second query into a large language model (LLM) (Somech para. 0003: the contextual data or metadata is input into a large language model (LLM)); and
returning an answer generated by the LLM as a response to the first user query (Somech para. 0003: the large language model (LLM) produces the answer: “Red Sea is due on Mar. 2 2023.”).
Somech does not appear to explicitly disclose a respective semantic similarity between the first user query and each element of the set of elements.
However, Kotaru teaches:
receiving at a question and answer (QA) system, a first user query (Kotaru para. Fig. 2 and para. 0054: user query 205);
obtaining, from one or more databases based at least in part on the first user query, a set of elements comprising one or more documents, one or more knowledge graph triplets, or both (Kotaru para. 0054: obtaining text samples from database);
assigning, based at least in part on a respective semantic similarity between the first user query and each element of the set of elements, a respective relevance score to each element of the set of elements (Kotaru para. 0067-0068: assigning similarity metrics for matching user query to most relevant database samples);
ranking the set of elements based at least in part on the respective relevance score assigned to each element of the set of elements (Kotaru para. 0068: ranking database samples based on similarity metric);
extracting additional information from each element of the set of elements based at least in part on the first user query (Kotaru para. 0069: most similar vectors in the database are extracted as context to feed into a large language model (LLM) along with the user query);
generating a second query based at least in part on the first user query and using the extracted additional information and the ranked set of elements (Kotaru para. 0069-0070: most similar vectors in the database are extracted as context to feed into a large language model (LLM) along with the user query);
inputting the generated second query into a large language model (LLM) (Kotaru para. 0069-0070: most similar vectors in the database are extracted as context to feed into a large language model (LLM) along with the user query); and
returning an answer generated by the LLM as a response to the first user query (Kotaru para. Fig. 2 and para. 0054: response 210).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to have modified Somech to include the teachings of Kotaru because it allows efficient retrieval of the most relevant database data to provide context for a user’s query (Kotaru para. 0067-0070).
As to claim 7, Somech as modified by Kotaru teaches wherein, to input the generated second query into the LLM, the one or more processors are individually or collectively operable to execute the code to cause the apparatus to:
generate a response template comprising the generated second query, the first user query, the extracted additional information, and the ranked set of elements;, wherein the response template includes formatting elements (Kotaru para. 0070 and Table 2: prompt template input to LLM); and
input the response template into the LLM, wherein the LLM generates the answer based on the response template (Kotaru para. 0070 and Table 2: prompt template input to LLM).
As to claim 8, Somech teaches an apparatus, comprising:
one or more memories storing processor-executable code (Somech para. 0135 and Fig. 11: processor 14 executing instructions); and
one or more processors coupled with the one or more memories and individually or collectively operable to execute the code to cause the apparatus to (Somech para. 0135 and Fig. 11: memory 12 storing instructions for processor 14):
receiving at a question and answer (QA) system, a first user query (Somech para. 0003: a user issues a query, such as “when is Red Sea due?”);
obtaining, from one or more databases based at least in part on the first user query, a set of elements comprising one or more documents, one or more knowledge graph triplets, or both (Somech para. 0003: corpus data supplement, e.g. a meeting transcript document);
assigning a respective relevance score to each element of the set of elements (Somech para. 0053: ranking based on relevance score);
ranking the set of elements based at least in part on the respective relevance score assigned to each element of the set of elements (Somech para. 0053: ranking based on relevance score);
extracting additional information from each element of the set of elements based at least in part on the first user query (Somech para. 0003: the corpus data supplement provides contextual data or metadata);
generating a second query based at least in part on the first user query and using the extracted additional information and the ranked set of elements (Somech para. 0003: the contextual data or metadata is input into a large language model (LLM));
inputting the generated second query into a large language model (LLM) (Somech para. 0003: the contextual data or metadata is input into a large language model (LLM)); and
returning an answer generated by the LLM as a response to the first user query (Somech para. 0003: the large language model (LLM) produces the answer: “Red Sea is due on Mar. 2 2023.”).
Somech does not appear to explicitly disclose a respective semantic similarity between the first user query and each element of the set of elements.
However, Kotaru teaches:
receiving at a question and answer (QA) system, a first user query (Kotaru para. Fig. 2 and para. 0054: user query 205);
obtaining, from one or more databases based at least in part on the first user query, a set of elements comprising one or more documents, one or more knowledge graph triplets, or both (Kotaru para. 0054: obtaining text samples from database);
assigning, based at least in part on a respective semantic similarity between the first user query and each element of the set of elements, a respective relevance score to each element of the set of elements (Kotaru para. 0067-0068: assigning similarity metrics for matching user query to most relevant database samples);
ranking the set of elements based at least in part on the respective relevance score assigned to each element of the set of elements (Kotaru para. 0068: ranking database samples based on similarity metric);
extracting additional information from each element of the set of elements based at least in part on the first user query (Kotaru para. 0069: most similar vectors in the database are extracted as context to feed into a large language model (LLM) along with the user query);
generating a second query based at least in part on the first user query and using the extracted additional information and the ranked set of elements (Kotaru para. 0069-0070: most similar vectors in the database are extracted as context to feed into a large language model (LLM) along with the user query);
inputting the generated second query into a large language model (LLM) (Kotaru para. 0069-0070: most similar vectors in the database are extracted as context to feed into a large language model (LLM) along with the user query); and
returning an answer generated by the LLM as a response to the first user query (Kotaru para. Fig. 2 and para. 0054: response 210).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to have modified Somech to include the teachings of Kotaru because it allows efficient retrieval of the most relevant database data to provide context for a user’s query (Kotaru para. 0067-0070).
As to claim 14, see the rejection of claim 7 above.
As to claim 15, Somech teaches a non-transitory computer-readable medium storing code, the code comprising instructions executable by one or more processors to (Somech para. 0135 and Fig. 11: memory 12 storing instructions for processor 14):
one or more memories storing processor-executable code (Somech para. 0135 and Fig. 11: processor 14 executing instructions); and
one or more processors coupled with the one or more memories and individually or collectively operable to execute the code to cause the apparatus to (Somech para. 0135 and Fig. 11: memory 12 storing instructions for processor 14):
receiving at a question and answer (QA) system, a first user query (Somech para. 0003: a user issues a query, such as “when is Red Sea due?”);
obtaining, from one or more databases based at least in part on the first user query, a set of elements comprising one or more documents, one or more knowledge graph triplets, or both (Somech para. 0003: corpus data supplement, e.g. a meeting transcript document);
assigning a respective relevance score to each element of the set of elements (Somech para. 0053: ranking based on relevance score);
ranking the set of elements based at least in part on the respective relevance score assigned to each element of the set of elements (Somech para. 0053: ranking based on relevance score);
extracting additional information from each element of the set of elements based at least in part on the first user query (Somech para. 0003: the corpus data supplement provides contextual data or metadata);
generating a second query based at least in part on the first user query and using the extracted additional information and the ranked set of elements (Somech para. 0003: the contextual data or metadata is input into a large language model (LLM));
inputting the generated second query into a large language model (LLM) (Somech para. 0003: the contextual data or metadata is input into a large language model (LLM)); and
returning an answer generated by the LLM as a response to the first user query (Somech para. 0003: the large language model (LLM) produces the answer: “Red Sea is due on Mar. 2 2023.”).
Somech does not appear to explicitly disclose a respective semantic similarity between the first user query and each element of the set of elements.
However, Kotaru teaches:
receiving at a question and answer (QA) system, a first user query (Kotaru para. Fig. 2 and para. 0054: user query 205);
obtaining, from one or more databases based at least in part on the first user query, a set of elements comprising one or more documents, one or more knowledge graph triplets, or both (Kotaru para. 0054: obtaining text samples from database);
assigning, based at least in part on a respective semantic similarity between the first user query and each element of the set of elements, a respective relevance score to each element of the set of elements (Kotaru para. 0067-0068: assigning similarity metrics for matching user query to most relevant database samples);
ranking the set of elements based at least in part on the respective relevance score assigned to each element of the set of elements (Kotaru para. 0068: ranking database samples based on similarity metric);
extracting additional information from each element of the set of elements based at least in part on the first user query (Kotaru para. 0069: most similar vectors in the database are extracted as context to feed into a large language model (LLM) along with the user query);
generating a second query based at least in part on the first user query and using the extracted additional information and the ranked set of elements (Kotaru para. 0069-0070: most similar vectors in the database are extracted as context to feed into a large language model (LLM) along with the user query);
inputting the generated second query into a large language model (LLM) (Kotaru para. 0069-0070: most similar vectors in the database are extracted as context to feed into a large language model (LLM) along with the user query); and
returning an answer generated by the LLM as a response to the first user query (Kotaru para. Fig. 2 and para. 0054: response 210).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to have modified Somech to include the teachings of Kotaru because it allows efficient retrieval of the most relevant database data to provide context for a user’s query (Kotaru para. 0067-0070).
Claims 2-4, 9-11, and 16-18 are rejected under 35 U.S.C. 103 as being unpatentable over Somech and Kotaru as applied to claims 1, 8, and 15 above, and further in view of Gupta et al. (U.S. Patent Application Publication No. 20220083552 A1, hereinafter referred to as Gupta).
As to claim 2, Somech as modified by Kotaru does not appear to explicitly disclose wherein the QA system comprises a directed acyclic graph (DAG) structure configured to generate the second query based on the first user query.
However, Gupta teaches wherein the QA system comprises a directed acyclic graph (DAG) structure configured to generate the second query based on the first user query (Gupta para. 0043, 0047, and 0089: DAG structure configured for query rewriting/transformation).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to have modified Somech as modified by Kotaru to include the teachings of Gupta because it allows for optimizing queries at runtime (Gupta para. 0043).
As to claim 3, Somech as modified by Kotaru and Gupta teaches wherein the QA system operates in accordance with a smart mode or thoughtful mode, and the DAG structure is dependent on whether the QA system is in the smart mode or the thoughtful mode, (Gupta para. 0084 and Table 14: DAG structure dependent upon execution mode).
As to claim 4, Somech as modified by Kotaru and Gupta teaches wherein:
when operating in accordance with the smart mode, the DAG structure includes a first set of nodes (Gupta para. 0084 and Table 14: DAG nodes dependent upon execution mode), and
when operating in accordance with the thoughtful mode, the DAG structure includes a second set of nodes (Gupta para. 0084 and Table 14: DAG nodes dependent upon execution mode).
As to claim 9, see the rejection of claim 2 above.
As to claim 10, see the rejection of claim 3 above.
As to claim 11, see the rejection of claim 4 above.
As to claim 16, see the rejection of claim 2 above.
As to claim 17, see the rejection of claim 3 above.
As to claim 18, see the rejection of claim 4 above.
Claims 5, 12, and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Somech and Kotaru as applied to claims 1, 8, and 15 above, and further in view of McElvain, Gayle (U.S. Patent Application Publication No. 20250124066 A1, hereinafter referred to as McElvain).
As to claim 5, Somech as modified by Kotaru does not appear to explicitly disclose wherein inputting the generated second query into the LLM results in a rewritten query, further comprising: obtaining, from the one or more databases and based at least in part on the rewritten query, a second set of elements comprising one or more second documents, one or more second knowledge graph triplets, or both; assigning, based at least in part on a respective semantic similarity between the rewritten query and each element of the second set of elements, a respective relevance score to each element of the second set of elements; ranking the second set of elements based at least in part on the respective relevance score assigned to each element of the second set of elements; and extracting second additional information from each element of the second set of elements based at least in part on the rewritten query, wherein the second additional information is input to the LLM and wherein the answer generated by the LLM is based at least in part on the second information.
However, McElvain teaches wherein inputting the generated second query into the LLM results in a rewritten query (McElvain para. 0042: LLM used to generate multiple reformulations of query), further comprising:
obtaining, from the one or more databases and based at least in part on the rewritten query, a second set of elements comprising one or more second documents (McElvain para. 0031: obtaining additional information from documents), one or more second knowledge graph triplets, or both;
assigning, based at least in part on a respective semantic similarity between the rewritten query and each element of the second set of elements, a respective relevance score to each element of the second set of elements (McElvain para. 0044 and 0046: relevance scoring based on semantic similarity);
ranking the second set of elements based at least in part on the respective relevance score assigned to each element of the second set of elements (McElvain para. 0044 and 0046: relevance scoring and ranking based on semantic similarity); and
extracting second additional information from each element of the second set of elements based at least in part on the rewritten query, wherein the second additional information is input to the LLM and wherein the answer generated by the LLM is based at least in part on the second information (McElvain abstract: LLM generates textual content asosciated with search criteria).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to have modified Somech as modified by Kotaru to include the teachings of McElvain because it enables query expansion, which improves recall for complex questions (McElvain para. 0042).
As to claim 12, see the rejection of claim 5 above.
As to claim 19, see the rejection of claim 5 above.
Claims 6, 13, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Somech and Kotaru as applied to claims 1, 8, and 15 above, and further in view of Jeon et al. (U.S. Patent Application Publication No. 20260050801 A1, hereinafter referred to as Jeon).
As to claim 6, Somech as modified by Kotaru does not appear to explicitly disclose wherein: the one or more knowledge graph triplets are obtained from one or more knowledge graphs generated by the LLM from a corpus of documents, and each knowledge graph triplet comprises a subject element, a relationship element, and an object element.
However, Jeon teaches wherein:
the one or more knowledge graph triplets are obtained from one or more knowledge graphs generated by the LLM from a corpus of documents (Jeon para. 0058, 0076, and Fig. 1: knowledge triplet generation module 105 uses an LLM to generate knowledge triplets from document database 20), and
each knowledge graph triplet comprises a subject element, a relationship element, and an object element (Jeon Fig. 5: knowledge triplets in the form: (subject, relationship, object)).
It would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to have modified Somech as modified by Kotaru to include the teachings of Jeon because use of an LLM provides the ability to understand and generate a vast amount of text data (Jeon para. 0055).
As to claim 13, see the rejection of claim 6 above.
As to claim 20, see the rejection of claim 6 above.
Contact Information
Any inquiry concerning this communication or earlier communications from the examiner should be directed to UMAR MIAN whose telephone number is (571)270-3970. The examiner can normally be reached Monday to Friday, 10 am to 6:30 pm.
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, Tony Mahmoudi can be reached on (571) 272-4078. 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.
/Umar Mian/
Examiner, Art Unit 2163
1 MPEP § 2106.07(a)(III)(A) explains that a specification demonstrates the well-understood, routine, conventional nature of additional elements when it describes the additional elements as well-understood or routine or conventional ( or an equivalent term) or in a manner that indicates that the additional elements are sufficiently well-known that the specification does not need to describe the particulars of such additional elements to satisfy 35 U.S.C. § 112(a).