DETAILED ACTION
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA and is in response to communications filed on 6/14/2024 in which claims 1-20 are presented for examination.
Priority
Acknowledgment is made of applicant’s claim for foreign priority under 35 U.S.C. 119 (a)-(d). The certified copy has been filed in parent Application No. 63/509,218, filed on 6/20/2023.
Considerations under - 35 USC § 101
Claims 1-20 are NOT directed to an abstract idea because at least the following is true: the claims recite a practical application in the form of improvement to machine learning by providing a database dictionary to reduce hallucinations of AI production of query code.
Dependent claims also recite significantly more than a judicial exception due at least to their dependency on their respective independent claims.
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 1-3, 7, 11-13, 17 are rejected under 35 U.S.C. 103 as being unpatentable over Chao et al. US 20240403373 A1 (hereinafter referred to as “Chao”) in view of Amulu et al. US 20240184793 A1 (hereinafter referred to as “Amulu”).
As per claim 1, Chao teaches:
A system for generating query code in response to a natural language input, the system comprising:
at least one processor operatively connected to a memory, the processor when executing configured to:
accept a request comprising a natural language input (Chao, [0115] – A search query 502 is received. The search query 502 may be a series of words (e.g., “weather”) and, in some embodiments, may be a natural-language statement);
associate the natural language input with a database dictionary definition comprising metadata defining architecture of a database target to query (Chao, [0155] – The data dictionary schema 1500 includes a high-level summary 1502 of the shared data provided by the data listing 423, table information 1504 describing the tables included in the shared data, view information 1506 describing views in the shared data, function information 1508 describing functions in the shared data and stored procedure information 1510 describing stored procedures in the shared data. Per column information 1512 is also provided for each table and view in the shared data. As shown, each of the objects includes a set of data fields. For example, the high-level summary information 1502 includes a number of schemas, tables, views, functions and stored procedures in the shared data. [0167] – The results insight further comprises a related query based on the search query and the plurality of data listings. [0168] – The data dictionary for each of the plurality of data listings comprising metadata describing data shared by the data listing and metadata describing individual objects included in the data shared by the data listing);
input a combination of the natural language and database dictionary definition into a large language model configured to produce query code (Chao, [0101] – The generative engine 514 of the LLM 510 may be used to provide conversational-like output and/or reasoning about the data at several stages of the process. This may include explaining the data, explaining the relevance of results to the user, and/or dynamically providing the user with example SQL queries for the selected data listings 423. [0115] – Natural language input. [0155] – The data dictionary schema 1500 includes a high-level summary 1502 of the shared data provided by the data listing 423, table information 1504 describing the tables included in the shared data, view information 1506 describing views in the shared data, function information 1508 describing functions in the shared data and stored procedure information 1510 describing stored procedures in the shared data. Per column information 1512 is also provided for each table and view in the shared data. As shown, each of the objects includes a set of data fields. For example, the high-level summary information 1502 includes a number of schemas, tables, views, functions and stored procedures in the shared data. Fig. 5 also shows the LLM used to both ingest and provide data to and from the search engine which processes the search query and data listings);
capture a query code output produced from the combination, the output tailored to the request (Chao, [0106] – The embedding engine 512 may examine the data listing 423D, including the contents of the database referenced by the data listing 423D (including the database schema(s)) and/or metadata of the data listing 423D. [0137] – The query 1110 provided to the generative engine 514 may be of the form “Given a SQL table with columns: <data listing table columns> and another table with columns <user table columns> provide a SQL query using both tables.” As illustrated in FIG. 11B, the generative engine 514 may automatically (e.g., through the use of the LLM 510) respond with a response 1120. In the example of FIG. 11B, the response 1120 includes both an example SQL query as well as an explanation of the various parts of the SQL query, including its output),
the database dictionary definition (Chao, [0155] – The data dictionary schema 1500 includes a high-level summary 1502 of the shared data provided by the data listing 423, table information 1504 describing the tables included in the shared data), and
…
display the query code output (Chao, [0125] – The ranking results may be provided as part of user interface that is generated, which may be displayed to the user who provided the search query).
Although Chao teaches databases where data is retrieved, Chao doesn’t explicitly teach a target database, however, Amulu teaches:
the target database (Amulu, [0021] – Target); and
It would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to modify Chao’s invention in view of Amulu in order to utilize an internal routing aspect (identifying one or more targets for corresponding database queries; this is advantageous for finding one specific record, e.g. a part in a manufacturer's database, and can be sufficient to search by Part Number in the obvious location of a Parts table (Amulu, paragraph [0024]).
As per claim 2, Chao as modified teaches:
The system of claim 1, wherein the query code output produced from the combination is executable on the target database (Amulu, [0021] – A database query handler (sometimes called a “database engine”) has a front-end aspect, namely receiving a client query; an internal routing aspect (identifying one or more targets for corresponding database queries, e.g. determining where to search); and a back-end aspect, namely planning and executing a database query on each target).
As per claim 3, Chao as modified teaches:
The system of claim 1, wherein the system is configured to access a predefined database dictionary definition responsive to selection in a user interface, wherein the predefined database dictionary definition is predefined relative to processing the natural language input (Amulu, [0052] – A facility of an unstructured data source that provides organization information is dubbed a “map.” and is a counterpart to the schema, dictionaries, and descriptions of a database. That is, the map can provide definition, description, and relationships between organizational entities within the unstructured data source. Whereas the schema, dictionaries, and descriptions can be regarded as data objects, a map can incorporate data objects or methods, in any combination. To illustrate, MONGODB® offers a method “db.inventory.find({ })” to obtain all documents in a collection, with many variants for narrower data retrieval).
As per claim 7, Chao as modified teaches:
The system of claim 1, wherein the at least one processor is configured to fine-tune the large language model by the input of database dictionary definition (Chao, [0087]-[0089] – Training the LLM).
Claims 11-13, 17 are directed to a method performing steps recited in claims 1-3, 7 with substantially the same limitations. Therefore, the rejections made to claims 1-3, 7 are applied to claims 11-13, 17.
Claims 4 and 14 are rejected under 35 U.S.C. 103 as being unpatentable over Chao in view of Amulu and further in view of Parikh et al. US 20160283503 A1 (hereinafter referred to as “Parikh”).
As per claim 4, Chao as modified doesn’t explicitly teach that a manual update can be given to a database dictionary, however, Parikh teaches:
The system of claim 1, wherein the system is configured to update the database dictionary definition responsive to manual input in a user interface (Parikh, [0024] – A user, via the user interface discussed previously, can change, edit or modify any aspects of databases, tablespaces, partitions, bitmaps, database dictionaries, or database tables at any time or during any step of workflow).
It would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to modify Chao’s invention in view of Parikh in order to allow a user to change any aspects of databases, database dictionaries, etc.; this is advantageous for scaling of an organization and general flexibility (Parikh, paragraph [0024]).
Claim 14 is directed to a method performing steps recited in claim 4 with substantially the same limitations. Therefore, the rejections made to claim 4 are applied to claim 14.
Claims 5-6, 8-10, 15-16, and 18-20 are rejected under 35 U.S.C. 103 as being unpatentable over Chao in view of Amulu and further in view of Rahmfeld et al. US 20210390099 A1 (hereinafter referred to as “Rahmfeld”).
As per claim 5, Chao as modified teaches:
The system of claim 1, wherein the at least one processor is configured to:
…
automatically generate at least a portion of the database dictionary definition (Chao, [0156] – The user interface 1550 includes a data dictionary 1528 generated for the listing).
Chao as modified doesn’t explicitly teach an acceptance to the database schema information, however, Rahmfeld teaches:
accept database schema information (Rahmfeld, [0007] – Determining, based on the utterance and the zero or more previous data conversation steps, an effective schema targeted by the utterance); and
It would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to modify Chao’s invention in view of Rahmfeld in order to accept a schema; this is advantageous because the system can select the right schema based on the utterance (Rahmfeld, paragraph [0032]).
As per claim 6, Chao as modified with Rahmfeld teaches:
The system of claim 5, wherein the automatically generated code includes at least one of a table group definition, join definition, attribute definition, phrase definition, alias definition, or look up definition, union definition, or comment definition (Chao, [0045] – Joins).
As per claim 8, Rahmfeld as modified teaches:
The system of claim 1, wherein the at least one processor is configured to:
accept information for a database architecture for a target database (Rahmfeld, [0018] – The (translated or directly submitted) structured query language query is executed against the targeted one or more data sets which produces a data result set and metadata for a given data conversation step);
execute a plurality of rules on the information for the database architecture (Rahmfeld, [0007] – Executing the executable structured query language statement for the data query engine schema);
automatically define table groups based on relationships established in the target database or a database schema associated with the target database based on execution of the plurality of rules (Rahmfeld, [0030] – Multiple entities can be grouped within a topic. The tables supporting such entities should have well-defined join conditions. [0033] – Joined tables); and
store a database dictionary including at least the table groups for optimizing query code generation (Chao, [0044] – Within each file, the values of each attribute or column are grouped together and compressed using a scheme sometimes referred to as hybrid columnar. Each table has a header which, among other metadata, contains the offsets of each column within the file.),
defined independent of submission of a respective request comprising the natural language input (Amulu, [0052] – A facility of an unstructured data source that provides organization information is dubbed a “map.” and is a counterpart to the schema, dictionaries, and descriptions of a database. That is, the map can provide definition, description, and relationships between organizational entities within the unstructured data source. Whereas the schema, dictionaries, and descriptions can be regarded as data objects, a map can incorporate data objects or methods, in any combination. To illustrate, MONGODB® offers a method “db.inventory.find({ })” to obtain all documents in a collection, with many variants for narrower data retrieval).
As per claim 9, Rahmfeld as modified teaches:
The system of claim 8, wherein the at least one processor is configured to automatically define at least one of a join definition, an attribute definition, a phrase definition, an alias definition, a look up definition, union definition, or a comment definition based on execution of the plurality of rules (Rahmfeld, [0024] – Join conditions are interpreted as join definitions. [0054] – Translation module 130 may generate an executable structured query language statement based on the intermediate structured query language statement including derivation and insertion of From/Join and/or Group-By expressions in the executable structured query language statement).
As per claim 10, Rahmfeld as modified teaches:
The system of claim 1, wherein the at least one processor is configured to:
accept information for a database architecture for a target database (Rahmfeld, [0018] – The (translated or directly submitted) structured query language query is executed against the targeted one or more data sets which produces a data result set and metadata for a given data conversation step);
execute at least a plurality of rules on the information for the database architecture (Rahmfeld, [0007] – Executing the executable structured query language statement for the data query engine schema); and
automatically define at least one of a table group, a join definition, an attribute definition, a phrase definition, an alias definition, a look up definition, a union definition, or a comment definition based on execution of the plurality of rules (Rahmfeld, [0024] – Join conditions are interpreted as join definitions. [0054] – Translation module 130 may generate an executable structured query language statement based on the intermediate structured query language statement including derivation and insertion of From/Join and/or Group-By expressions in the executable structured query language statement); and
store a database dictionary including at least the table groups for optimizing query code generation (Chao, [0156] – The data dictionary 1528 allows a data consumer to view information describing a group of objects included in the shared data, such as tables, views and functions).
Claims 15-16,18-20 are directed to a method performing steps recited in claims 5-6, 8-10 with substantially the same limitations. Therefore, the rejections made to claims 5-6, 8-10 are applied to claims 15-16,18-20.
Response to Arguments
The judicial exception rejection of the claims has been withdrawn due to the amendments made to the claims and arguments made in Remarks of 11/11/2025.
Applicant’s arguments with respect to claims have been considered but are generally 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.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
Kate et al. “Conversion of Natural Language Query to SQL Query”, 2018 Second international conference on electronics, communication and aerospace technology (ICECA).
Lober, et al. US 9176996 B2 teaches: in response to a determination that the table definition for the table in a source database dictionary is not compatible with a table definition for the table in a target database dictionary, the INSERT/SELECT statement execution is rolled back in response to a determination that the data from the first INSERT/SELECT statement did not fit in a table in the target schema of the target database (Abstract).
Bui et al. US 20230306061 A1 teaches: automatically generating and analyzing database queries. In various embodiments, a database inquiry assistance system maintains a first machine learning model trained using query history data for a database and a second machine learning model using analysis history for the database.
Yadav et al. US 20210019309 A1 teaches mapping natural language to queries using a query grammar (Abstract).
Bhatt et al. US 20170308524 A1.
Leary et al. US 12380282 B2.
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.
Contact Information
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Matthew Ellis whose telephone number is (571)270-3443. The examiner can normally be reached on Monday-Friday 8AM-5PM.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Neveen Abel-Jalil can be reached on (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 an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
February 5, 2026
/MATTHEW J ELLIS/Primary Examiner, Art Unit 2152