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 .
Response to Arguments
Applicant’s arguments filed 9 July 2026 have been fully considered.
Applicant argues that generating the claimed first structured data query and generating the claimed second structured data query do not recite abstract ideas because they are “specific computer-centric interactions.” Examiner respectfully disagrees. The generating steps are recited at a high level of generality. MPEP § 2106.04(a)(2)(III)(C)(3); see Berkheimer v. HP, Inc., 881 F.3d 1360, 1366 (Fed. Cir. 2018). The claims do not recite specific computer functionality because they fail to recite any limitation as to how the LLM generates the respective structured data queries but merely invoke the LLM as a generic tool to perform the mental process of forming judgments, at a high level of generality, as to the contents of the respective data queries.
Applicant argues that the claim represents an improvement to “generation and execution of structured database queries.” Examiner respectfully disagrees. Generation of structured database queries is not a technology, but a mental process; i.e., using knowledge of a query language, humans write structured database queries in order to translate a desire to request answers expressed in natural language into queries that can be executed by a computer, at a high level of generality, on a knowledge base. Those claim elements directed to execution of the queries are not analyzed under step 2A, prong one.
Applicant argues that the claims integrate the recited abstract idea into a practical application. Examiner respectfully disagrees. Generating specific structured database queries is a desired outcome, and not a particular way to achieve the desired outcome. MPEP § 2106.05(a); see McRO, Inc. v. Bandai Namco Games Am. Inc., 837 F.3d 1299, 1314-15 (Fed. Cir. 2016). The claims’ execution of the generated structured database queries is insignificant extra-solution activity as generic data retrieval.
Applicant argues that, as an ordered combination, the invention improves technology by providing additional context which allows the LLM to more accurately create a structured query. Examiner respectfully disagrees. As an ordered combination, the invention is to a) receive, without limitation as to mechanism of receiving, a query; b) form a judgment as to the contents of a structured query based on the question and a database schema, where the judgment is formed by a LLM predicting the content without limitation as to the mechanism of prediction; c) form a second judgment as to whether existing values in a data store validate the prediction based on semantic similarity, where the values in the data store are searched without limitation as to the mechanism of searching, and semantic similarity is judged without limitation as to the mechanism of judging similarity; d) form a third judgment as to the contents of a second structured query based on the judgment of validation; e) querying the structured database using the structured database, without limitation as to how the database is queried; and f) outputting the results on an interface, without limitation as to how the interface outputs the results.
Providing additional context also allows a human to more accurately create a structured query. MPEP § 2106.05(f)(3). As stated in the grounds of rejection:
I can guess that the SQL query “SELECT COUNT(CallOutcome = ‘Open Account’) FROM Calls WHERE Caller=’TeamA’” would answer the natural language question “How many calls by team A open accounts?” and then look at the database to figure out what the text literals for CallOutcome, Open Account, Calls, Caller, and TeamA should actually be, in order to refine the query to give the correct answer (e.g., because I realize, upon observing the database, that the database uses ‘Create Account’ not ‘Open Account’ as the value to represent the semantic meaning that an account was opened as a result of a call, that the query should read “SELECT COUNT(CallOutcome = ‘Create Account’) FROM Calls WHERE Caller=’TeamA’”).
The recited process of predicting a first structured data query, validating the first structured data query against existing values in a data stored, and updating the first structured data query to a corrected second structured database query based on the validation therefore has general applicability.
There is nothing in the claims that explains how the LLM uses the provided context to more accurately create a structured query. MPEP § 2106.05(f)(1). The recitation that the first and third judgments are performed by an LLM are not a meaningful limitations, and as an ordered combination the invention is mere automation of the preexisting mental process of translating a natural language query to a structured query language query by, at a high level of generality, predicting schema values, validating the predictions, and then correcting errors based on the validation, all the while using a LLM as a mere tool.
Applicant argues that Sun does not teach validating a first predicted text value extracted from preliminary SQL queries. Examiner respectfully disagrees. Sun generates a first structured data query based on a natural language question using conventional Text-to-SQL, and “applies complete prompt (schema plus auxiliary information) [sic],” Sun pg. 15, to the preliminary SQL to obtain refined SQL. In the example given, based on the query “What is the highest eligible free rate for K-12 students in the schools in Alameda County?”, conventional Text-to-SQL produces the inferred SQL “SELECT ‘FRPM Count (K-12)’ / ‘Enrollment (K-12)’ FROM frpm WHERE ‘County Name’ = ‘Alameda County’ ORDER BY(CAST(‘FRPM Count (K-12)’ AS REAL) / ‘Enrollment (K-12)’) DESC LIMIT 1;”. Sun pp. 15, 23. Sun then “appli[es] complete prompt (schema plus auxiliary information) to obtain accurate SQL,’ Sun pg. 15, and having validated that the database contains “Alameda” and not “Alameda County,” replaces “Alameda County” in the preliminary SQL with “Alameda” in the refined SQL. Sun pg. 23. Sun therefore generates the refined query based on ii) the natural language query “What is the highest eligible free rate for K-12 students in the schools in Alameda County?” ii) the schema, by incorporating the table name and column names from the schema iii), existing values, by validating that the database contains a value “Alameda” and not “Alameda County”, and iv) the first structured data query, by replacing “Alameda County” with “Alameda” based on the validation. Applicant does not explain how applying complete prompt (schema plus auxiliary information) to preliminary SQL to obtain more accurate SQL, as explicitly recited, Sun pg. 15, differs from the claimed invention.
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 per claims 1, 10 and 19:
The claim(s) recites an abstract idea.
The limitation, “generating, using the LLM based on the first natural language question and a database schema for the structured database, a first structured data query having a first predicted text value determined by the LLM based on the first natural language question,” as drafted, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components. For example, in the context of this limitation, “generating” encompasses a person forming a judgment what structured query would, when executed, provide an answer to the natural language question by guessing as to what semantics are used by the database to represent the semantic meaning underlying the question, e.g., that a SQL query seeking to answer the question “How many calls by team A open accounts?” would select on a column named CallType for a value “Open Account”. This limitation therefore falls within the “Mental Processes” grouping of abstract ideas. MPEP § 2106.04(a)(2)(III).
The limitation, “searching a data store for one or more existing values from the discrete text values that validate the first predicted text value extracted from the first structured data query, wherein the data store comprises semantically similar values for the discrete text values, and wherein the searching identifies the one or more existing values from the discrete text values based on a similarity threshold of the semantically similar values to the first predicted text value,” as drafted, is a process that, under its broadest reasonable interpretation, covers a calculation (i.e., calculating similarity and calculating a Boolean comparison to the threshold), and covers performance of the limitation in the mind but for the recitation of generic computer components. For example, in the context of this limitation, “searching” encompasses a person forming a judgment as to, given calculated similarity scores for the semantically similar discrete text values, which score is above a threshold. This limitation therefore falls within the “Mathematical Concepts” and “Mental Processes” groupings of abstract ideas. MPEP §§ 2106.04(a)(2)(I), 2106.04(a)(2)(III).
The limitation, “generating, using the LLM based on (i) the first natural language question, (ii) the database schema, (iii) the one or more existing values from the discrete text values, and (iv) the first structured data query, a second structured data query for querying the structured database for the structured data, wherein the second structured data query refines the first structured data query by including a second predicted text value based on the one or more existing values for querying the structured database,” as drafted, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components. For example, in the context of this limitation, “generating” encompasses a person forming a judgment that, based on the schema, existing values, and guessed query, that the actual semantics of the database specify that the query should select on a column named CallType for a value “Create Account” instead of “Open Account”. This limitation therefore falls within the “Mental Processes” grouping of abstract ideas. MPEP § 2106.04(a)(2)(III).
I can guess that the SQL query “SELECT COUNT(CallOutcome = ‘Open Account’) FROM Calls WHERE Caller=’TeamA’” would answer the natural language question “How many calls by team A open accounts?” and then look at the database to figure out what the text literals for CallOutcome, Open Account, Calls, Caller, and TeamA should actually be, in order to refine the query to give the correct answer (e.g., because I realize, upon observing the database, that the database uses ‘Create Account’ not ‘Open Account’ as the value to represent the semantic meaning that an account was opened as a result of a call, that the query should read “SELECT COUNT(CallOutcome = ‘Create Account’) FROM Calls WHERE Caller=’TeamA’”). That the mental steps of guessing and refining are recited as being performed using a generic LLM still recites a mental process. MPEP § 2106.04(a)(2)(III)(C)(3).
Accordingly, the claim(s) recites abstract ideas. MPEP § 2106.04(a). These abstract ideas can be considered together as a single abstract idea for the purposes of Step 2A Prong Two and Step 2B, MPEP § 2106.04(II)(B), namely a process of writing a structured query on a database to answer a natural language question by guessing a structured query containing text literals specifying fields and values, comparing those fields and values in the guess to fields and values in the database, and refining the text literals in the guess based on the comparison.
The abstract idea of writing a structured query to answer a natural language question is not integrated into a practical application.
The additional element, “receiving, via a user interface of an application, a first natural language question for structured data stored by a structured database, wherein the structured data includes one or more columns having discrete text values,” is insignificant extra-solution activity as mere data gathering. MPEP § 2106.05(g).
The additional element, “querying the structured database using the second structured data query for a response to the first natural language question based at least on the structured data,” is insignificant extra-solution activity as mere data gathering. MPEP § 2106.05(g).
The additional element, “outputting, in the user interface of the application, the response to the first natural language question,” is insignificant extra-solution activity as mere data gathering. MPEP § 2106.05(g).
As an ordered combination, the invention merely automates the existing manual process of using a structured query to answer a natural language question, using generic LLM computer functionality as a tool to perform a mental process. MPEP §§ 2106.05(a), 2106.05(f).
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claim(s) as a whole, and therefore the claim is directed to an abstract idea. MPEP § 2106.04(d).
As discussed above with respect to integration of the abstract idea into a practical application, the conclusions for the additional elements being generic computer components and mere instructions to apply on a computer, insignificant extra-solution activity, and/or mere field of use limitations are carried over and these additional elements do not provide significantly more than the abstract idea. MPEP § 2106.05(II).
In re-evaluating the limitations that are insignificant extra-solution activity, the following limitations represent elements that have been recognized as well-understood, routine, conventional activity within the field of computer functions:
The additional element, “receiving, via a user interface of an application, a first natural language question for structured data stored by a structured database, wherein the structured data includes one or more columns having discrete text values,” is well-understood, routine, and conventional activity because it is collecting a response to presented information that is recited at a high level of generality similar to the activity of using a computer interface to collect user input. MPEP §§ 2106.07(a)(III)(B), 2106.05(d)(II); see OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363-64 (Fed. Cir. 2015).
The additional element, “querying the structured database using the second structured data query for a response to the first natural language question based at least on the structured data,” is well-understood, routine, and conventional activity because querying using a SQL query is described, Specification [0004], as conventional. MPEP § 2106.07(a)(III)(A).
The additional element, “outputting, in the user interface of the application, the response to the first natural language question,” is well-understood, routine, and conventional activity because it is presenting information in a manner that is recited at a high level of generality similar to the activity of using a computer interface to present information. MPEP §§ 2106.07(a)(III)(B), 2106.05(d)(II); see OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363-64 (Fed. Cir. 2015).
As an ordered combination, the claim simply appends well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the abstract idea of writing a structured query to answer a natural language question because the claim as a whole amounts to nothing more than generic computer functions merely used to implement the abstract idea. MPEP §§ 2106.07(a)(III)(B), 2106.05(d)(II); see BASCOM Global Internet Servs. v. AT&T Mobility LLC, 827 F.3d 1341, 1349 (Fed. Cir. 2016).
Accordingly, the claim(s) does not recite additional elements, either individually or in combination, that amount to significantly more than the abstract idea. MPEP § 2106.05. Therefore, as the claim(s) is directed to an abstract idea and does not recite additional elements that amount to significantly more than the abstract idea, the claim(s) is not patentable. MPEP § 2106.
As per claims 2, 11, and 20:
The claim(s) recites an abstract idea.
The limitation, “wherein the database schema comprises a list of data tables stored by the structured database, column identifiers of the one or more columns of the data tables, and descriptions for the data tables in the list,” as drafted, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components. For example, in the context of this limitation, “generating” encompasses a person forming a judgment that, based on a list of data tables stored by the structured database, column identifiers of the one or more columns of the data tables, descriptions for the data tables in the list, existing values, and guessed query, that the actual semantics of the database specify that the query should select on a column named CallType for a value “Create Account” instead of “Open Account”. This limitation therefore falls within the “Mental Processes” grouping of abstract ideas. MPEP § 2106.04(a)(2)(III).
Accordingly, the claim(s) recites an abstract idea. MPEP § 2106.04(a).
As the claim(s) recites no additional elements, the abstract idea is not integrated into a practical application, the claim is directed to the abstract idea, and the claim(s) does not amount to significantly more than the abstract idea. MPEP § 2106.07. Therefore, as the claim(s) is directed to an abstract idea and does not recite additional elements that amount to significantly more than the abstract idea, the claim(s) is not patentable. MPEP § 2106.
As per claims 3 and 12:
The claim(s) recite an abstract idea.
The limitation, “determining that the first structured data query includes a string literal corresponding to the first predicted text value,” as drafted, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components. For example, in the context of this limitation, “determining” encompasses a person forming a judgment that, e.g., “Open Account” is a string literal. This limitation therefore falls within the “Mental Processes” grouping of abstract ideas. MPEP § 2106.04(a)(2)(III).
The limitation, “wherein the data store comprises a vector database having vectors representing possible values of the discrete text values usable for queries to the structured database,” as drafted, is a process that, under its broadest reasonable interpretation, covers a calculation (i.e., calculating cosine similarity between two vectors). This limitation therefore falls within the “Mathematical Concepts” grouping of abstract ideas. MPEP § 2106.04(a)(2)(I).
Accordingly, the claim(s) recites abstract ideas. MPEP § 2106.04(a). These abstract ideas can be considered together as a single abstract idea, namely writing a structured query to answer a natural language question. MPEP § 2106.04(II)(B). This falls within the “Mental Processes” grouping of abstract ideas. MPEP § 2106.04(a)(2)(III).
As the claim(s) recites no additional elements, the abstract idea is not integrated into a practical application, the claim is directed to the abstract idea, and the claim(s) does not amount to significantly more than the abstract idea. MPEP § 2106.07. Therefore, as the claim(s) is directed to an abstract idea and does not recite additional elements that amount to significantly more than the abstract idea, the claim(s) is not patentable. MPEP § 2106.
As per claims 4 and 13:
The claim(s) recites an abstract idea.
The limitation, “extracting the string literal for the first predicted text value from the first structured data query,” as drafted, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components. For example, in the context of this limitation, “determining” encompasses a person forming a judgment that, e.g., that a vector embedding should be calculated for the string literal “Open Account”. This limitation therefore falls within the “Mental Processes” grouping of abstract ideas. MPEP § 2106.04(a)(2)(III).
The limitation, “converting the string literal to a first vector in a vector space,” as drafted, is a process that, under its broadest reasonable interpretation, covers a calculation (i.e., calculating a vector embedding). This limitation therefore falls within the “Mathematical Concepts” grouping of abstract ideas. MPEP § 2106.04(a)(2)(I).
The limitation, “comparing the first vector to the vectors representing the possible values in the vector space based on the similarity threshold,” as drafted, is a process that, under its broadest reasonable interpretation, covers a calculation (e.g., calculating cosine similarity). This limitation therefore falls within the “Mathematical Concepts” grouping of abstract ideas. MPEP § 2106.04(a)(2)(I).
Accordingly, the claim(s) recites abstract ideas. MPEP § 2106.04(a). These abstract ideas can be considered together as a single abstract idea, namely forming a judgment as to what string literal should be used in a SQL query to answer a natural language question based on a guess as to what the string literal in the query should be and existing discrete text values in the database being queried. MPEP § 2106.04(II)(B). This falls within the “Mental Processes” grouping of abstract ideas. MPEP § 2106.04(a)(2)(III).
As the claim(s) recites no additional elements, the abstract idea is not integrated into a practical application, the claim is directed to the abstract idea, and the claim(s) does not amount to significantly more than the abstract idea. MPEP § 2106.07. Therefore, as the claim(s) is directed to an abstract idea and does not recite additional elements that amount to significantly more than the abstract idea, the claim(s) is not patentable. MPEP § 2106.
As per claims 5 and 14:
The claim(s) recites an abstract idea.
The limitation, “wherein the similarity threshold utilizes a strict threshold and a lenient threshold, wherein the strict threshold causes the comparing to return a first list of n closest vectors of the vectors that are less than or equal to a first preset similarity value of the strict threshold, and wherein the lenient threshold causes the comparing to return to a second list of n closest vectors of the vectors that are less than or equal to a second preset similarity value that is a greater similarity distance than the first preset similarity value, and wherein the second present similarity value is configured to include related text values for the discrete text values,” as drafted, is a process that, under its broadest reasonable interpretation, covers a calculation. This limitation therefore falls within the “Mathematical Concepts” grouping of abstract ideas. MPEP § 2106.04(a)(2)(I).
Accordingly, the claim(s) recites an abstract idea. MPEP § 2106.04(a).
As the claim(s) recites no additional elements, the abstract idea is not integrated into a practical application, the claim is directed to the abstract idea, and the claim(s) does not amount to significantly more than the abstract idea. MPEP § 2106.07. Therefore, as the claim(s) is directed to an abstract idea and does not recite additional elements that amount to significantly more than the abstract idea, the claim(s) is not patentable. MPEP § 2106.
As per claims 6 and 15:
The claim(s) recites an abstract idea.
The limitation, “wherein the discrete text values comprise at least one of categorical observations, text identifiers, or text descriptions,” as drafted, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components. For example, in the context of this limitation, “generating” encompasses a person forming a judgment that, based on the schema, categorical observations, text identifiers, or text descriptions, and guessed query, that the actual semantics of the database specify that the query should select on a column named CallType for a value “Create Account” instead of “Open Account”. This limitation therefore falls within the “Mental Processes” grouping of abstract ideas. MPEP § 2106.04(a)(2)(III).
Accordingly, the claim(s) recites an abstract idea. MPEP § 2106.04(a).
As the claim(s) recites no additional elements, the abstract idea is not integrated into a practical application, the claim is directed to the abstract idea, and the claim(s) does not amount to significantly more than the abstract idea. MPEP § 2106.07. Therefore, as the claim(s) is directed to an abstract idea and does not recite additional elements that amount to significantly more than the abstract idea, the claim(s) is not patentable. MPEP § 2106.
As per claims 7 and 16:
The abstract idea of writing a structured query to answer a natural language question is not integrated into a practical application.
The additional element, “receiving a result to the second structured data query based on the querying, wherein the result identifies whether at least one data record having the first predicted text value or the second predicted text value is found based on the querying,” is insignificant extra-solution activity as mere data gathering. MPEP § 2106.05(g).
The additional element, “the response includes the result,” is insignificant extra-solution activity as mere data gathering. MPEP § 2106.05(g).
As an ordered combination, the invention merely automates the existing manual process of using a structured query to answer a natural language question, using generic LLM computer functionality as a tool to perform a mental process. MPEP §§ 2106.05(a), 2106.05(f).
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claim(s) as a whole, and therefore the claim is directed to an abstract idea. MPEP § 2106.04(d).
As discussed above with respect to integration of the abstract idea into a practical application, the conclusions for the additional elements being generic computer components and mere instructions to apply on a computer, insignificant extra-solution activity, and/or mere field of use limitations are carried over and these additional elements do not provide significantly more than the abstract idea. MPEP § 2106.05(II).
In re-evaluating the limitations that are insignificant extra-solution activity, the following limitations represent elements that have been recognized as well-understood, routine, conventional activity within the field of computer functions:
The additional element, “receiving a result to the second structured data query based on the querying, wherein the result identifies whether at least one data record having the first predicted text value or the second predicted text value is found based on the querying,” is well-understood, routine, and conventional activity because querying using a SQL query is described, Specification [0004], as conventional. MPEP § 2106.07(a)(III)(A).
The additional element, “the response includes the result,” is well-understood, routine, and conventional activity because it is presenting information in a manner that is recited at a high level of generality similar to the activity of using a computer interface to present information. MPEP §§ 2106.07(a)(III)(B), 2106.05(d)(II); see OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363-64 (Fed. Cir. 2015).
As an ordered combination, the claim simply appends well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the abstract idea of writing a structured query to answer a natural language question because the claim as a whole amounts to nothing more than generic computer functions merely used to implement the abstract idea. MPEP §§ 2106.07(a)(III)(B), 2106.05(d)(II); see BASCOM Global Internet Servs. v. AT&T Mobility LLC, 827 F.3d 1341, 1349 (Fed. Cir. 2016).
Accordingly, the claim(s) does not recite additional elements, either individually or in combination, that amount to significantly more than the abstract idea. MPEP § 2106.05. Therefore, as the claim(s) is directed to an abstract idea and does not recite additional elements that amount to significantly more than the abstract idea, the claim(s) is not patentable. MPEP § 2106.
As per claims 8 and 17:
The claim(s) recites an abstract idea.
The limitation, “wherein the result identifies that no data record was found based on the querying, and wherein the query generation operations further comprise: providing a list of the one or more existing values from the discrete text values and a suggestion of a second natural language question usable as an alternative to the first natural language question,” as drafted, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components. For example, in the context of this limitation, “providing” encompasses a person forming a judgment that, e.g., based on there not being results for a query, “SELECT COUNT(CallOutcome = ‘Create Account’) FROM Calls WHERE Caller=’TeamA’”, that ‘Create Account’ could be ‘Account Updates’, and to suggest “How many calls by Team A update accounts?” as a second natural language question. This limitation therefore falls within the “Mental Processes” grouping of abstract ideas. MPEP § 2106.04(a)(2)(III).
Accordingly, the claim(s) recites an abstract idea. MPEP § 2106.04(a).
As the claim(s) recites no additional elements, the abstract idea is not integrated into a practical application, the claim is directed to the abstract idea, and the claim(s) does not amount to significantly more than the abstract idea. MPEP § 2106.07. Therefore, as the claim(s) is directed to an abstract idea and does not recite additional elements that amount to significantly more than the abstract idea, the claim(s) is not patentable. MPEP § 2106.
As per claims 9 and 18:
The abstract idea of writing a structured query to answer a natural language question is not integrated into a practical application.
The additional element, “wherein the generating the first and the second structured data queries using the LLM includes prompting the LLM using a prompt having instructions to generate a structured data query from the first natural language question using at least the database schema and based on a knowledge base of a structured query language and a structured query language syntax,” is mere instruction to use a computer as a tool to generate a structured data query from the first natural language question using at least the database schema and based on a knowledge base of a structured query language and a structured query language syntax because it invokes the LLM as a tool to perform an existing process of query generation. MPEP § 2106.05(f).
As an ordered combination, the invention merely automates the existing manual process of using a structured query to answer a natural language question, using generic LLM computer functionality as a tool to perform a mental process. MPEP §§ 2106.05(a), 2106.05(f).
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claim(s) as a whole, and therefore the claim is directed to an abstract idea. MPEP § 2106.04(d).
As discussed above with respect to integration of the abstract idea into a practical application, the conclusions for the additional elements being generic computer components and mere instructions to apply on a computer, insignificant extra-solution activity, and/or mere field of use limitations are carried over and these additional elements do not provide significantly more than the abstract idea. MPEP § 2106.05(II).
In re-evaluating the limitations that are mere instructions to apply on a computer, the following limitations represent elements that have been recognized as well-understood, routine, conventional activity within the field of computer functions:
The additional element, “wherein the generating the first and the second structured data queries using the LLM includes prompting the LLM using a prompt having instructions to generate a structured data query from the first natural language question using at least the database schema and based on a knowledge base of a structured query language and a structured query language syntax,” is well-understood, routine, and conventional activity because it is described, Specification [0021] (“using GPT-4”), as a commercially available product. MPEP § 2106.07(a)(III)(A).
As an ordered combination, the claim simply appends well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the abstract idea of writing a structured query to answer a natural language question because the claim as a whole amounts to nothing more than generic computer functions merely used to implement the abstract idea. MPEP §§ 2106.07(a)(III)(B), 2106.05(d)(II); see BASCOM Global Internet Servs. v. AT&T Mobility LLC, 827 F.3d 1341, 1349 (Fed. Cir. 2016).
Accordingly, the claim(s) does not recite additional elements, either individually or in combination, that amount to significantly more than the abstract idea. MPEP § 2106.05. Therefore, as the claim(s) is directed to an abstract idea and does not recite additional elements that amount to significantly more than the abstract idea, the claim(s) is not patentable. MPEP § 2106.
Claim Rejections - 35 USC § 102
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claim(s) 1-20 is/are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Sun et al., SQL-PaLM: Improved large language model adaptation for Text-to-SQL (extended).
As per claims 1, 10, and 19, Sun teaches:
receiving, via a user interface of an application, a first natural language question for structured data stored by a structured database, wherein the structured data includes one or more columns having discrete text values, Sun pp. 13-16, 23, where, e.g., a question such as “What is the highest eligible free rate for K-12 students in the schools in Alameda County?” is received;
generating, using the LLM based on the first natural language question and a database schema for the structured database, a first structured data query having a first predicted text value determined by the LLM based on the first natural language question, Sun pg. 15 (“Using these limited inputs (only the schema), we can generate preliminary SQL queries for column selection.”);
searching a data store for one or more existing values from the discrete text values that validate the first predicted text value extracted from the first structured data query, wherein the data store comprises semantically similar values for the discrete text values, and wherein the searching identifies the one or more existing values from the discrete text values based on a similarity threshold of the semantically similar values to the first predicted text value, Sun pp. 13-15, where cosine similarity is used to obtain semantically similar database values;
generating, using the LLM based on (i) the first natural language question, (ii) the database schema, (iii) the one or more existing values from the discrete text values, and (iv) the first structured data query, a second structured data query for querying the structured database for the structured data, wherein the second structured data query refines the first structured data query based by including a second predicted text value based on the on the one or more existing values for querying the structured database, Sun pp. 13-15, 23, where, e.g., the text value ‘Alameda County’ replaces ‘Alameda’;
querying the structured database using the second structured data query for a response to the first natural language question based at least on the structured data, Sun pp. 16-17, where the generated SQL is executed; and
outputting, in the user interface of the application, the response to the first natural language question, Sun pp. 16-17, where the results of SQL are used to answer the question.
As per claims 2, 11, and 20, the rejection of claims 1, 10, and 19 is incorporated, and Sun further teaches:
wherein the database schema comprises a list of data tables stored by the structured database, column identifiers of the one or more columns of the data tables, and descriptions for the data tables in the list, Sun pg. 10, where the input X contains tables, columns of the tables, and descriptions.
As per claims 3 and 12, the rejection of claims 1 and 10 is incorporated, and Sun further teaches:
wherein the data store comprises a vector database having vectors representing possible values of the discrete text values usable for queries to the structured database, and wherein, before the searching, the query generation operations further comprise: determining that the first structured data query includes a string literal corresponding to the first predicted text value, Sun pg. 13 (“embedding similarity (e.g. the cosine distance between the embedding of wi and v(k)rj above a threshold is a match).”).
As per claims 4 and 13, the rejection of claims 3 and 12 is incorporated, and Sun further teaches:
extracting the string literal for the first predicted text value from the first structured data query; converting the string literal to a first vector in a vector space; and comparing the first vector to the vectors representing the possible values in the vector space based on the similarity threshold, Sun pg. 13 (“embedding similarity (e.g. the cosine distance between the embedding of wi and v(k)rj above a threshold is a match).”).
As per claims 5 and 14, the rejection of claims 4 and 13 is incorporated, and Sun further teaches:
wherein the similarity threshold utilizes a strict threshold and a lenient threshold, wherein the strict threshold causes the comparing to return a first list of n closest vectors of the vectors that are less than or equal to a first preset similarity value of the strict threshold, Sun pg. 13 (“Here, we use Fm as the longest contiguous matching subsequence approach (Cormen et al., 2022), as it allows us to accurately extract the exact values stored in the database.”), and wherein the lenient threshold causes the comparing to return to a second list of n closest vectors of the vectors that are less than or equal to a second preset similarity value that is a greater similarity distance than the first preset similarity value, and wherein the second present similarity value is configured to include related text values for the discrete text values, Sun pg. 13 (“embedding similarity (e.g. the cosine distance between the embedding of wi and v(k)rj above a threshold is a match).”).
As per claims 6 and 15, the rejection of claims 1 and 10 is incorporated, and Sun further teaches:
wherein the discrete text values comprise at least one of categorical observations, text identifiers, or text descriptions, Sun pp. 13-15, where the text values are identifiers.
As per claims 7 and 16, the rejection of claims 1 and 10 is incorporated, and Sun further teaches:
wherein, before outputting the response, the query generation operations further comprise: receiving a result to the second structured data query based on the querying, wherein the result identifies whether at least one data record having the first predicted text value or the second predicted text value is found based on the querying, wherein the response includes the result, Sun pg. 23, where “Alameda” is found but “Alameda County” is not, and the result reflects that “Alameda” is found.
As per claims 8 and 17, the rejection of claims 7 and 16 is incorporated, and Sun further teaches:
wherein the result identifies that no data record was found based on the querying, and wherein the query generation operations further comprise: providing a list of the one or more existing values from the discrete text values and a suggestion of a second natural language question usable as an alternative to the first natural language question, Sun pg. 23, where the existing value “Alameda” is provided, where the given error reason suggests the question “What is the highest eligible free rate for K-12 students in the schools in Alameda?” as an alternative.
As per claims 9 and 18, the rejection of claims 1 and 17 is incorporated, and Sun further teaches:
wherein the generating the first and the second structured data queries using the LLM includes prompting the LLM using a prompt having instructions to generate a structured data query from the first natural language question using at least the database schema and based on a knowledge base of a structured query language and a structured query language syntax, Sun pg. 15, where prompting is used to generate both the preliminary SQL and accurate SQL.
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.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to WILLIAM SPIELER whose telephone number is (571)270-3883. The examiner can normally be reached Monday-Friday, 11-3.
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, Ann Lo can be reached at 571-272-9767. 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.
WILLIAM SPIELER
Primary Examiner
Art Unit 2159
/WILLIAM SPIELER/ Primary Examiner, Art Unit 2159
Prosecution Insights