Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Detailed Action
This is an Office Action for application 19/060,999 in response to arguments and amendments filed on 12/02/2025. Claims 1, 3, 11, 13 and 20 are currently amended. Claims 2 and 12 are cancelled. Claims 1, 3-11 and 13-20 are pending and examined below.
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 12/02/2025 has been entered.
Response to Arguments
Applicant's other arguments filed 12/02/2025 have been fully considered but they are not persuasive.
Applicant first argues that newly amended claim 1 language that contains some language from cancelled claim 2
retrieving, by a data retrieval system on a computing device, a target data table set from the database based on the table query similarity, the target data table set comprising data associated with the user query, wherein retrieving the target data table set from the database comprises:
retrieving a first data table set from the database based on the query vector and the summary vector;
retrieving a second data table set from the database based on the query vector and the field vector; and
determining the target data table set based on the first data table set and the second data table set;
is not currently taught by the combination of Ranganathan and Hou because the previously cited portion of Ranganathan only teaches enhancing the model with semantic information, not retrieving a first data table set from the database based on the query vector and the summary vector as required by newly amended claim 1 language. Applicant also argues that the new language specifying that the query is a vector and that the summary is a vector is not taught by Ranganathan.
However, the cited portion of Ranganathan teaches that the grammar builder enhances the model with additional semantic information. While this is not the action of retrieval itself, this additional semantic information is the basis of the summary vector. As already explained in previous claim 1 limitation with respect to this cited section, the DB querier (#135) retrieves the data in response to the user query. Combined together, the query is retrieving the information from the table based on the semantic information.
Further, while Ranganathan is silent with respect to vectors and does not teach these new aspects of new claim 1 language, Hou does teach this new language. As shown in previously cited portions of Hou (Fig. 4; Abs., Pars. [0012, 48-9, 75]), the selection is done in an embedding space (i.e. using vectors) of a user query against a table summary and description of columns. This vector disclosure combined with Ranganathan’s query system is a query system that is capable of using vectors. In response to applicant's arguments against the references individually, one cannot show nonobviousness by attacking references individually where the rejections are based on combinations of references. See In re Keller, 642 F.2d 413, 208 USPQ 871 (CCPA 1981); In re Merck & Co., 800 F.2d 1091, 231 USPQ 375 (Fed. Cir. 1986). Therefore, argument is unpersuasive.
Applicant further argues that the new claim 1 limitation
retrieving a second data table set from the database based on the query vector and the field vector; and
is not taught by cited reference art Ranganathan because Ranganathan doesn’t disclose the vector aspects of the queries.
However, as explained above, Hou teaches this limitation, and piecemeal analysis of these reference arts is not an argument for nonobviousness. Therefore, argument is unpersuasive.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1, 3-4, 11, 13-14 and 20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Ranganathan et al. (US Pat. 11,604,790) in view of Hou et al. (US Pub. 2025/0061104).
Regarding claim 1, Ranganathan teaches
A computer-implemented method for data field and field retrieval, comprising: determining, by a data retrieval system on a computing device, a table query similarity between a user query and each data table in a database based on a query vector of the user query, a summary vector of a data table summary, and a field vector of a field name; (Fig. 1; Col. 10 [Line 62] - Col. 11 [Line 20], Col. 13 [Lines 5-20] a user's natural language (NL) query is translated into a SQL query that uses entity information for fields (i.e. field name) and enhanced semantic information (i.e. data summary) associated with the database (Examiner notes that the vector aspects are taught by Hou as shown below, but is being left here for referential clarity))
retrieving, by a data retrieval system on a computing device, a target data table set from the database based on the table query similarity, the target data table set comprising data associated with the user query, wherein retrieving the target data table set from the database comprises:
retrieving a first data table set from the database based on the query vector and the summary vector; (Fig. 1; Col. 8 [Line 59] - Col. 9 [Line 24] the grammar builder (#150) enhances the model with additional semantic information (i.e. data table summary) to help bridge a semantic gap between a physical data and a users' understanding of the domain and DB query, and the data is eventually retrieved by the DB querier (#135))
retrieving a second data table set from the database based on the query vector and the field vector; and (Col. 3 [Line 24 - Line 53] the translation from natural language to SQL can include a sequence of queries (i.e. a first and second query using different criteria))
determining the target data table set based on the first data table set and the second data table set; (Col. 3 [Line 24 - Line 53] the translation from natural language to SQL can include a sequence of queries (i.e. a first and second query using different criteria) to obtain query results (i.e. based on first and second dataset))
determining, by a data retrieval system on a computing device, a field query similarity between the user query and each field of each data table in the target data table set based on the user query and the field name; (Fig. 1; Col. 10 [Line 62] - Col. 11 [Line 20], Col. 13 [Lines 5-20] a user's natural language (NL) query is translated into a SQL query that uses entity information, relationships and values of attribute fields)
retrieving, by a data retrieval system on a computing device, a target field set from each data table in the target data table set based on the field query similarity; (Fig. 1; Col. 8 [Lines 59-63] the SQL query is by the DB querier (#135) used to retrieve data in response to the user query)
determining, by a data retrieval system on a computing device, a retrieval result of the data table and field retrieval based on the target data table set and the corresponding target field set; (Fig. 1; Col. 8 [Lines 59-63] the SQL query is by the DB querier (#135) used to retrieve data in response to the user query)
wherein the data retrieval result comprises the target data table set and the corresponding target field set; (Fig. 1; Col. 8 [Lines 59-63] the SQL query is by the DB querier (#135) used to retrieve data (i.e. retrieval result) in response to the user query)
causing, by the data table retrieval system, display the retrieval result on all computing devices. (Col. 11 Lines [11-20] query results are returned and displayed to a user)
Ranganathan does not explicitly teach
a table query similarity between a user query and each data table in a database based on a query vector of the user query, a summary vector of a data table summary, and a field vector of a field name;
However, from the same field, Hou teaches
a table query similarity between a user query and each data table in a database based on a query vector of the user query, a summary vector of a data table summary, and a field vector of a field name; (Fig. 4; Abs. Pars. [0012, 48-9, 75]; columns are selected based on a likelihood estimation in an embedding space (e.g. using vectors) between a user query and a table summary (i.e. summary vector of a data table summary) and description of columns (i.e. field vector of a field name))
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to combine the vector generation of Hou into the NL query system of Ranganathan. The motivation for this combination would have been to improve the speed of the system and overcome high computational needs as explained in Hou (Par. [0037]).
Regarding claim(s) 3, Ranganathan and Hou teach claim 1 as shown above, and Ranganathan further teaches
retrieving the second data table set from the database by determining a vector similarity between the query vector and the field vector; and (Col. 3 [Line 24 - Line 53] the translation from natural language to SQL can include a sequence of queries (i.e. a first and second query using different criteria))
determining the target data table set by performing deduplication and fusion on the first data table set and the second data table set. (Fig. 1; Col. 10 [Line 54] - Col. 11 [Line 10] a consolidated (i.e. deduplicated and fused) comparison of sales figures is presented to the user)
Hou further teaches
The method of claim 2, wherein retrieving the target data table set from the database comprises: generating the query vector of the user query through a pre-trained model; (Fig. 4; Par. [0012, 72] the user's query (#408) is given a corresponding vector also referred to as a question vector (i.e. query vector))
obtaining the summary vector of the data table summary from a vector library, the summary vector being generated through the pre-trained model; (Fig. 4, Fig. 7; Par. [0012, 48, 72] during table summary step (#402) an LLM generates a summary of a table and a column and are eventually turned into associated vector in a vector database (#710))
retrieving the first data table set from the database by determining a vector similarity between the query vector and the summary vector; (Fig. 4, Fig. 6; Par. [0012, 48, 72] an LLM generates a summary of a table and a column and are eventually turned into associated vector in a vector database, and the user query is performed on the vector database (i.e. similarity between the query and summary vector))
obtaining the field vector of the field name from the vector library, the field vector being generated through the pre-trained model; (Fig. 4, Fig. 6; Par. [0012, 48, 72] an LLM generates a summary of a table and a column (i.e. field) and are eventually turned into associated vector (e.g. via a pre-trained model) in a vector database, and the user query is performed on the vector database)
Regarding claim(s) 4, Ranganathan and Hou teach claim 1 as shown above, and Ranganathan further teaches
The method of claim 1, wherein retrieving the target field set from each data table in the target data table set comprises: retrieving a first field set from each data table in the target data table set based on the user query and the field name; (Fig. 1; Col. 10 [Line 62] - Col. 11 [Line 20], Col. 13 [Lines 5-20] a user's natural language (NL) query is translated into a SQL query that uses entity information for fields (i.e. field name))
determining the target field set of each data table in the target data table set based on the first field set and the second field set. (Col. 3 [Line 24 - Line 53] the translation from natural language to SQL can include a sequence of queries (i.e. a first and second query using different criteria) to obtain query results (i.e. based on first and second dataset))
Hou further teaches
retrieving the second field set from each data table in the target data table set based on the user query and a field vector of the field name; and (Fig. 4; Par. [0012, 72] the user's query (#408) is given a corresponding vector also referred to as a question vector (i.e. query vector))
Regarding claim 11, while worded slightly differently than claim 1, is rejected under a similar rationale. Ranganathan further teaches
a processor; and a memory coupled to the processor (Col. 29 [Lines 54-65] a processor and memory are used to implement the system)
Regarding claim 13, while worded slightly differently than claim 3, is rejected under a similar rationale.
Regarding claim 14, while worded slightly differently than claim 4, is rejected under a similar rationale.
Regarding claim 20, while worded slightly differently than claim 11, is rejected under a similar rationale.
Claim(s) 5-7 and 15-17 is/are rejected under 35 U.S.C. 103 as being unpatentable over Ranganathan et al. (US Pat. 11,604,790) in view of Hou et al. (US Pub. 2025/0061104), and further in view of Das et al. (US Pub. 2019/0034429).
Regarding claim(s) 5, Ranganathan and Hou teach claim 1 as shown above, and Ranganathan further teaches
retrieving the first field set from each data table in the target data table set by determining a literal similarity between the rewritten user query and the field name; (Fig. 1; Col. 10 [Line 62] - Col. 11 [Line 20], Col. 13 [Lines 5-20] a user's natural language (NL) query is translated into a SQL query that uses entity information for fields (i.e. field name))
retrieving the second field set from each data table in the target data table set by determining a vector similarity between the query vector and the field vector; and (Fig. 1; Col. 10 [Line 62] - Col. 11 [Line 20], Col. 13 [Lines 5-20] a user's natural language (NL) query is translated into a SQL query that uses entity information for fields (i.e. field name))
determining the target field set by performing deduplication and fusion on the first field set and the second field set based on a predefined rule. (Fig. 1; Col. 10 [Line 54] - Col. 11 [Line 10] a consolidated (i.e. deduplicated and fused) comparison of sales figures is presented to the user)
Hou further teaches
generating the query vector of the user query through a pre-trained model; (Fig. 4, Fig. 6; Par. [0012, 48, 72] an LLM generates a summary of a table and a column (i.e. field) and are eventually turned into associated vector (e.g. via a pre-trained model) in a vector database, and the user query is performed on the vector database)
The combination of Ranganathan and Hou do not explicitly teach
The method of claim 4, wherein retrieving the target field set from each data table in the target data table set comprises: generating a rewritten user query based on domain knowledge associated with the user query;
However, from the same field, Das teaches
The method of claim 4, wherein retrieving the target field set from each data table in the target data table set comprises: generating a rewritten user query based on domain knowledge associated with the user query; (Fig. 9; Par. [0220] the natural language (NL) request (#915) is matched to an intent by the request processing engine (#920) informed by a data scope engine (#910) connected to domain-specific data (i.e. based on domain knowledge; #820) and generates (i.e. rewrites) an unambiguated NL request)
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to combine the domain-specific knowledge of Das into the NL query system of Ranganathan. The motivation for this combination would have been to improve the effectiveness of NL applications as explained in Das (Par. [0237]).
Regarding claim(s) 6, Ranganathan, Hou and Das teach claim 5 as shown above, and Ranganathan further teaches
The method of claim 5, wherein retrieving the second field set from each data table in the target data table set further comprises: determining a field heat based on field statistical data; and (Col. 12 [Line 46] - Col. 13 [Line 4] heat maps are calculated in post-processed results)
Hou further teaches
retrieving the second field set based on the vector similarity and the field heat. (Fig. 4, Fig. 6; Par. [0012, 48, 72] an LLM generates a summary of a table and a column (i.e. field) and are eventually turned into associated vector in a vector database, and the user query is performed on the vector database)
Regarding claim(s) 7, Ranganathan and Hou teach claim 1 as shown above, but do not explicitly teach
The method of claim 1, further comprising: obtaining domain knowledge associated with the user query; and
generating a rewritten user query based on the user query and the domain knowledge.
However, from the same field, Das teaches
The method of claim 1, further comprising: obtaining domain knowledge associated with the user query; and (Fig. 9; Par. [0220] the natural language (NL) request (i.e. user query; #915) is matched to an intent by the request processing engine (#920) informed by a data scope engine (#910) connected to domain-specific data (i.e. based on domain knowledge; #820) and generates (i.e. rewrites) an unambiguated NL request)
generating a rewritten user query based on the user query and the domain knowledge. (Fig. 9; Par. [0220] the natural language (NL) request (#915) is matched to an intent by the request processing engine (#920) informed by a data scope engine (#910) connected to domain-specific data (i.e. based on domain knowledge; #820) and generates (i.e. rewrites) an unambiguated NL request)
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to combine the domain-specific knowledge of Das into the NL query system of Ranganathan. The motivation for this combination would have been to improve the effectiveness of NL applications as explained in Das (Par. [0237]).
Regarding claim 15, while worded slightly differently than claim 5, is rejected under a similar rationale.
Regarding claim 16, while worded slightly differently than claim 6, is rejected under a similar rationale.
Regarding claim 17, while worded slightly differently than claim 7, is rejected under a similar rationale.
Claim(s) 8-10 and 18-19 is/are rejected under 35 U.S.C. 103 as being unpatentable over Ranganathan et al. (US Pat. 11,604,790) and Hou et al. (US Pub. 2025/0061104), and further in view of Lai (US Pub. 2020/0302122).
Regarding claim(s) 8, Ranganathan and Hou teach claim 1 as shown above, and Ranganathan further teaches
The method of claim 1, further comprising: generating prompt information based on the user query and the target data table set; (Fig. 1; Col. 8 [Lines 26-39] the generated natural language expression is based on the user's expression and is based on the schema and values of the database)
Ranganathan and Hou do not explicitly teach
generating a semantic ranking score of the target data table set using a pre-trained model based on the prompt information; and
retrieving a third data table set from the target data table set based on the semantic ranking score.
However, from the same field, Lai teaches
generating a semantic ranking score of the target data table set using a pre-trained model based on the prompt information; and (Par. [0072, 84] results are scored and ranked using machine learning models and presented to the user)
retrieving a third data table set from the target data table set based on the semantic ranking score. (Par. [0072, 84] results are scored and ranked using machine learning models and presented to the user)
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to combine the ranking system of Lai into the NL query system of Ranganathan. The motivation for this combination would have been to allow users to ask questions without having technical knowledge of the underlaying data as explained in Lai (Par. [0030]).
Regarding claim(s) 9, Ranganathan, Hou and Lai teach claim 8 as shown above, and Ranganathan further teaches
The method of claim 8, further comprising: determining a heat ranking score of the third data table set based on table statistical data of each data table in the third data table set; and (Col. 12 [Line 46] - Col. 13 [Line 4] heat maps are calculated in post-processed results)
Lai further teaches
retrieving a fourth data table set from the third data table set based on the heat ranking score. (Par. [0072, 84] results are scored and ranked using machine learning models and presented to the user)
Regarding claim(s) 10, Ranganathan, Hou and Lai teach claim 9 as shown above, and Ranganathan further teaches
The method of claim 9, further comprising: obtaining a date field set of each data table in the fourth data table set; (Col. 21 [Line 49] - Col. 22 [Line 18] in filtering, date and trend criteria can be used for producing the final results)
determining a popular field set of each data table in the fourth data table set based on the field statistical data; and (Col. 21 [Line 49] - Col. 22 [Line 18] in filtering, date and trend (i.e. popular) criteria can be used for producing the final results)
generating a combined field set based on the date field set, the popular field set, and the target field set. (Col. 21 [Line 49] - Col. 22 [Line 18] in filtering, date and trend criteria can be used for producing the final results (i.e. the combined field data set))
Regarding claim 18, while worded slightly differently than claim 8, is rejected under a similar rationale.
Regarding claim 19, while worded slightly differently than claim 9, is rejected under a similar rationale.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to J MITCHELL CURRAN whose telephone number is (469)295-9081. The examiner can normally be reached M-F 8:00am - 5:00pm.
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, Sherief Badawi can be reached at (571) 272-9782. 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.
/J MITCHELL CURRAN/Examiner, Art Unit 2161
/YU ZHAO/Primary Examiner, Art Unit 2169