DETAILED ACTION
This office action is in response to the communication filed on April 30, 2026. Claims 1-20 are currently pending.
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 on April 30, 2026 have been fully considered but they are not persuasive for the following reasons:
Applicant in Pages 9-13 of the Remarks argues that the amended claims do not recite an abstract idea and, even if determined to recited an abstract idea, clearly integrate the abstract idea into a practical application.
Examiner respectfully disagrees.
It is important to note that the judicial exception alone cannot provide the improvement. The improvement can be provided by one or more additional elements (MPEP 2106.05(a)).
If the specification sets forth an improvement in technology, the claim must be evaluated to ensure that the claim itself reflects the disclosed improvement (MPEP 2106.04(d)(1)).
Amended independent claim 1 and similarly amended independent claims 11 and 20 cover several steps, such as the filtering and identifying steps, that recite an abstract idea within the “Mental Processes” grouping of abstract ideas, because a person can mentally or using a pen and paper perform the limitations recited in said steps, which is discussed in detail in the current 101 rejection below.
The remaining steps in the claims that are identified as reciting additional elements, such as the retrieving, sending, receiving, and obtaining steps, are only adding insignificant extra-solution activity to the judicial exception, are recognized as a well understood, routine, and conventional activity within the field of computer functions, and are applying the exception using generic computer components, which is not sufficient to amount to significantly more than the judicial exception and are not directed to any specific improvement in computer technology.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Applicant in Pages 13-16 of the Remarks argues that the cited prior art Mishaeli and Dines do not show or suggest the features in amended independent claims 1, 11, and 20.
Applicant’s arguments with respect to claim(s) 1-20 have been considered but are moot in view of new grounds of rejection, as discussed in detail in the 103 rejection below.
For the above reasons, Examiner states that rejection of the current Office action is proper.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-20 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 1 recites the limitation "the source table" in lines 9-10. There is insufficient antecedent basis for this limitation in the claim.
Claim 1 recites the limitation "the first database" in line 10. There is insufficient antecedent basis for this limitation in the claim.
Claim 1 recites the limitation "the plurality of target tables" in lines 10-11. There is insufficient antecedent basis for this limitation in the claim.
Claim 1 recites the limitation "the second database" in line 11. There is insufficient antecedent basis for this limitation in the claim.
Claim 1 recites the limitation "the schemas of the source table" in line 16. There is insufficient antecedent basis for this limitation in the claim.
Claim 11 recites the limitation "the source table" in line 4. There is insufficient antecedent basis for this limitation in the claim.
Claim 11 recites the limitation "the first database" in line 5. There is insufficient antecedent basis for this limitation in the claim.
Claim 11 recites the limitation "the plurality of target tables" in line 5. There is insufficient antecedent basis for this limitation in the claim.
Claim 11 recites the limitation "the second database" in line 6. There is insufficient antecedent basis for this limitation in the claim.
Claim 11 recites the limitation "the schemas of the source table" in line 11. There is insufficient antecedent basis for this limitation in the claim.
Claim 20 recites the limitation "the source table" in line 5. There is insufficient antecedent basis for this limitation in the claim.
Claim 20 recites the limitation "the first database" in line 6. There is insufficient antecedent basis for this limitation in the claim.
Claim 20 recites the limitation "the plurality of target tables" in line 6. There is insufficient antecedent basis for this limitation in the claim.
Claim 20 recites the limitation "the second database" in line 7. There is insufficient antecedent basis for this limitation in the claim.
Claim 20 recites the limitation "the schemas of the source table" in line 12. There is insufficient antecedent basis for this limitation in the claim.
Dependent claims 2-10 and 12-19 inherit the same deficiencies of their base claims, therefore, they are also indefinite.
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.
At step 1:
Independent claims 1, 11, and 20 respectively recite a computing system, a computer-implemented method, and one or more non-transitory computer-readable media, which are directed to a statutory category such as a process, machine, or an article of manufacture.
At step 2A, prong one:
Independent claim 1 and similarly independent claims 11 and 20 recite the limitations:
performing, during a first stage of a two-stage process:
“filtering, in runtime, the set of metadata based, at least in part, on semantic context to generate a reduced set of metadata”;
A person can mentally or using a pen and paper perform, during a first stage of a two-stage process, filtering, in runtime, of a set of metadata based at least in part on semantic context to generate a reduced set of metadata.
“identifying, in runtime, one or more matching tables among the plurality of target tables based on filtered metadata, wherein the identifying comprises constructing a first prompt based on the schemas of the source table and the schema of the plurality of target tables…”;
A person can mentally or using a pen and paper perform, during a first stage of a two-stage process, identifying, in runtime, of one or more matching tables among a plurality of target tables based on based on filtered metadata, wherein the identifying comprises mentally or using a pen and paper constructing a first prompt based on the schemas of the source table and the schema of the plurality of target tables.
performing, during a second stage of the two-stage process:
“identifying, in runtime, one or more pairs of matching attributes between the source table and the selected matching table based on comparison of the first sample attribute data and the second sample attribute data…”.
A person can mentally or using a pen and paper perform, during a second stage of a two-stage process, identifying, in runtime, of one or more pairs of matching attributes between a source table and a selected matching table based on comparison of a first sample attribute data and a second sample attribute data.
The limitations, as recited above, are processes that, under their broadest reasonable interpretation, cover steps that can be performed in the human mind or by a human using a pen and paper, but for recitation of generic computer components.
If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components, then it falls within the “Mental Processes” grouping of abstract ideas.
Accordingly, the claim recites an abstract idea.
At step 2A, prong two:
This judicial exception is not integrated into a practical application.
Independent claim 1 and similarly independent claims 11 and 20 recite the limitations:
performing, during a first stage of a two-stage process:
“retrieving, in runtime, a set of metadata comprising a schema of the source table from the first database and a schema of the plurality of target tables from the second database”, which is a step of performing retrieving of data. The step is recited at a high level of generality, and amounts to mere data gathering, which is a form of insignificant extra-solution activity.
“and sending the first prompt to a large language model”, which is a step of performing sending of data. The step is recited at a high level of generality, and amounts to mere data gathering, which is a form of insignificant extra-solution activity.
“receiving a response from the large language model comprising one or more matching tables based on first vector embeddings from the schema of the source table and second vector embeddings from the schema for the plurality of target tables meeting at least a threshold similarity metric”, which is a step of performing receiving of data. The step is recited at a high level of generality, and amounts to mere data gathering, which is a form of insignificant extra-solution activity.
performing during a second stage of the two-stage process:
“obtaining, in runtime, first sample attribute data from the source table and second sample attribute data from a selected matching table”, which is a step of performing obtaining or retrieving of data. The step is recited at a high level of generality, and amounts to mere data gathering, which is a form of insignificant extra-solution activity.
The additional elements “a computing system for improved schema matching of two databases, the computing system comprising: memory; one or more hardware processors coupled to the memory; and one or more non-transitory computer readable storage media storing instructions that, when loaded into the memory, cause the one or more hardware processors to perform operations comprising:”, “from the first database”, “from the second database”, “to a large language model”, “from the large language model”, and “using the large language model” in the steps in claim 1 are recited at a high-level of generality, such that it amounts to no more than mere instructions to apply the exception using generic computer components.
The additional elements “a computer-implemented method for improved schema matching of two databases, the method comprising:”, from the first database”, “from the second database”, “to a large language model”, “from the large language model”, and “using the large language model” in the steps in claim 11 are recited at a high-level of generality, such that it amounts to no more than mere instructions to apply the exception using generic computer components.
The additional elements “one or more non-transitory computer-readable media having encoded thereon computer-executable instructions causing one or more processors to perform a method for improved schema matching of two databases, the method comprising:”, from the first database”, “from the second database”, “to a large language model”, “from the large language model”, and “using the large language model” in the steps in claim 20 are recited at a high-level of generality, such that it amounts to no more than mere instructions to apply the exception using generic computer components.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claim a whole, because it does not impose any meaningful limits on practicing the abstract idea.
At step 2B:
Independent claims 1, 11, and 20 recite the same additional elements as identified in step 2A prong two above. These additional elements are not sufficient to amount to significantly more than the judicial exception.
Independent claim 1 and similarly independent claims 11 and 20 recite the limitations:
performing, during a first stage of a two-stage process:
“retrieving, in runtime, a set of metadata comprising a schema of the source table from the first database and a schema of the plurality of target tables from the second database”, which is a step of retrieving data, and is recognized as a well understood, routine, and conventional activity within the field of computer functions as an element of storing and retrieving information in memory (MPEP 2106.05(d)(II)(iv)).
“and sending the first prompt to a large language model”, which is a step of performing sending of data, and is recognized as a well understood, routine, and conventional activity within the field of computer functions as an element of receiving or transmitting data over a network (MPEP 2106.05(d)(II)(i)).
“receiving a response from the large language model comprising one or more matching tables based on first vector embeddings from the schema of the source table and second vector embeddings from the schema for the plurality of target tables meeting at least a threshold similarity metric”, which is a step of performing receiving of data, and is recognized as a well understood, routine, and conventional activity within the field of computer functions as an element of receiving or transmitting data over a network (MPEP 2106.05(d)(II)(i)).
performing during a second stage of the two-stage process:
“obtaining, in runtime, first sample attribute data from the source table and second sample attribute data from a selected matching table”, which is a step of obtaining or retrieving data, and is recognized as a well understood, routine, and conventional activity within the field of computer functions as an element of storing and retrieving information in memory (MPEP 2106.05(d)(II)(iv)).
Accordingly, the additional limitations are not sufficient to amount to significantly more than the judicial exception. Therefore, the claim is directed to an abstract idea and is not patent eligible.
Dependent claim 2 and similarly dependent claim 12 recites additional limitations, such as:
wherein identifying the one or more matching tables comprises:
“constructing, in runtime, the first prompt, wherein constructing the first prompt comprises inserting the schema of the source table and the schema of the plurality of target tables into a first prompt template”;
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 11, because a person can mentally or using a pen and paper construct, in runtime, a first prompt, wherein constructing the first prompt comprises mentally or using a pen and paper inserting a schema of the source table and a schema of plurality of target tables into a first prompt template, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
“prompting, in runtime, the large language model using the first prompt”, which is a step of merely applying the prompt, such that it amounts to no more than mere instructions to apply the exception using generic computer components.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Dependent claim 3 recites additional limitations, such as:
“wherein constructing the first prompt further comprises removing some attributes in the schema of the source table or the schema of the plurality of target tables from the first prompt based on one or more predefined filtering criteria”.
These limitations are directed to the same abstract idea under the mental processes grouping as independent claim 1, because a person can mentally or using a pen and paper construct a first prompt by mentally or using a pen and paper removing some attributes in a schema of the source table or schema of a plurality of target tables from the first prompt based on one or more predefined filtering criteria, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Dependent claim 4 and similarly dependent claim 13 recites additional limitations, such as:
“wherein retrieving the schema of the source table comprises obtaining, in runtime, text descriptions of the source table and attributes of the source table from a first dictionary associated with the first database,
wherein retrieving the schema of the plurality of target tables comprises obtaining, in runtime, text descriptions of the plurality of target tables and attributes of the plurality of target tables from a second dictionary associated with the second database”, which are steps of retrieving data.
At step 2A prong two, the step is recited at a high level of generality, and amounts to mere data gathering, which is a form of insignificant extra-solution activity.
At step 2B, the step is recognized as a well understood, routine, and conventional activity within the field of computer functions as an element of storing and retrieving information in memory (MPEP 2106.05(d)(II)(iv)).
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Dependent claim 5 and similarly dependent claim 14 recites additional limitations, such as:
“wherein retrieving the schema of the source table comprises obtaining, in runtime, statistics of the source table,
wherein retrieving the schema of the plurality of target tables comprises obtaining, in runtime, statistics of the plurality of target tables”, which are steps of retrieving data.
At step 2A prong two, the step is recited at a high level of generality, and amounts to mere data gathering, which is a form of insignificant extra-solution activity.
At step 2B, the step is recognized as a well understood, routine, and conventional activity within the field of computer functions as an element of storing and retrieving information in memory (MPEP 2106.05(d)(II)(iv)).
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Dependent claim 6 and similarly dependent claim 15 recites additional limitations, such as:
wherein identifying the one or more matching tables comprises:
“generating a first vector embedding based on the schema of the source table; and
generating second vector embeddings based on the schema of the plurality of target tables”.
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 11, because a person can mentally or using a pen and paper identify one or more matching tables by mentally or using a pen and paper generating a first vector embedding based on a schema of a source table and second vector embeddings based on schema of a plurality of target tables, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Dependent claim 7 recites additional limitations, such as:
wherein identifying the one or more matching tables further comprises:
“measuring table similarities between the first vector embedding and the second vector embeddings; and
identifying table similarities that are greater than a predefined threshold”.
These limitations are directed to the same abstract idea under the mental processes grouping as independent claim 1, because a person can mentally or using a pen and paper identify one or more matching tables by mentally or using a pen and paper measuring table similarities between a first vector embedding and second vector embeddings and mentally or using a pen and paper identifying table similarities that are greater than a predefined threshold, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Dependent claim 8 and similarly dependent claim 17 recites additional limitations, such as:
wherein identify one or more pairs of matching attributes comprises:
“constructing, in runtime, a second prompt”;
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 11, because a person can mentally or using a pen and paper identify one or more pairs of matching attributes by mentally or using a pen and paper constructing, in runtime, a second prompt, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
“prompting, in runtime, the large language model using the second prompt”, which is a step of merely applying the prompt, such that it amounts to no more than mere instructions to apply the exception using generic computer components.
“wherein constructing the second prompt comprises inserting the first sample attribute data and second sample attribute data into a second prompt template”.
These limitations are directed to the same abstract idea under the mental processes grouping as independent claims 1 and 11, because a person can mentally or using a pen and paper identify one or more pairs of matching attributes by mentally or using a pen and paper constructing a second prompt by inserting a first sample attribute data and second sample attribute data into the second prompt template”, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Dependent claim 9 recites additional limitations, such as:
“wherein constructing the second prompt further comprises removing some of the first sample attribute data or the second sample attribute data from the second prompt based on one or more predefined filtering criteria”.
These limitations are directed to the same abstract idea under the mental processes grouping as independent claim 1, because a person can mentally or using a pen and paper construct a second prompt by mentally or using a pen and paper removing some of a first sample attribute data or a second sample attribute data from the second prompt based on one or more predefined filtering criteria, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Dependent claim 10 recites additional limitations, such as:
wherein identify one or more pairs of matching attributes comprises:
“generating third vector embeddings based on the first sample attribute data;
generating fourth vector embeddings based on the second sample attribute data; and
measuring attribute similarities between the third vector embeddings and the fourth vector embeddings”.
These limitations are directed to the same abstract idea under the mental processes grouping as independent claim 1, because a person can mentally or using a pen and paper identify one or more pairs of matching attributes by mentally or using a pen and paper generating third vector embeddings based on a first sample attribute data, generating fourth vector embeddings based on a second sample attribute data, and measuring attribute similarities between the third vector embeddings and the fourth vector embeddings, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Dependent claim 16 recites additional limitations, such as:
wherein identifying the one or more matching tables further comprises:
“measuring table similarities between the first vector embedding and the second vector embeddings; and
ranking the one or more matching tables based on the table similarities”.
These limitations are directed to the same abstract idea under the mental processes grouping as independent claim 11, because a person can mentally or using a pen and paper identify one or more matching tables by mentally or using a pen and paper measuring table similarities between a first vector embedding and second vector embeddings and ranking one or more matching tables based on the table similarities, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Dependent claim 18 recites additional limitations, such as:
wherein identify one or more pairs of matching attributes comprises:
“generating third vector embeddings based on the first sample attribute data; and
generating fourth vector embeddings based on the second sample attribute data”.
These limitations are directed to the same abstract idea under the mental processes grouping as independent claim 11, because a person can mentally or using a pen and paper identify one or more pairs of matching attributes by mentally or using a pen and paper generating third vector embeddings based on a first sample attribute data and generating fourth vector embeddings based on a second sample attribute data, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Dependent claim 19 recites additional limitations, such as:
wherein identify one or more pairs of matching attributes further comprises:
“measuring attribute similarities between the third vector embeddings and the fourth vector embeddings; and
ranking the one or more pairs of matching attributes based on the attribute similarities”.
These limitations are directed to the same abstract idea under the mental processes grouping as independent claim 11, because a person can mentally or using a pen and paper identify one or more pairs of matching attributes by mentally or using a pen and paper measuring attribute similarities between third vector embeddings and fourth vector embeddings and ranking one or more pairs of matching attributes based on the attribute similarities, and because the limitations do not recite any additional elements that are sufficient to amount to significantly more.
Accordingly, the additional elements, individually or in combination, do not integrate the abstract idea into a practical application, even viewing the claims a whole, because it does not impose any meaningful limits on practicing the abstract idea.
Accordingly, dependent claims 2-10 and 12-19 are also directed to an abstract idea without significantly more and are not patent eligible.
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) 1-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Mishaeli (US Pub 2025/0278434, provisional application filing date 02/29/24) in view of Shah (US Pub 2025/0156419) in view of Klein (US Pub 2025/0307238) and in further view of Dines (US Pub 2023/0419161).
With respect to claim 1, Mishaeli discloses a computing system for improved schema matching of two databases (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema; Mishaeli in [0050] and [0052] discloses a computing system), the computing system comprising:
memory (Mishaeli in [0050] and [0052] discloses a computing system comprising memory);
one or more hardware processors coupled to the memory (Mishaeli in [0052] and [0055] discloses one or more processors coupled to memory; Mishaeli in [0059] and [0066] discloses hardware embodiment, implementations by hardware-based systems that perform specified functions or acts or combinations of special purpose hardware and computer instructions); and
one or more non-transitory computer readable storage media storing instructions that, when loaded into the memory, cause the one or more hardware processors to perform operations (Mishaeli in [0052] and [0063] discloses stored instructions executed by one or more processors and one or more memory included within computer system Mishaeli in [0059] and [0060] discloses computer program product on a computer-usable storage medium having computer-usable program code embodied in the medium, medium containing, storing, communicating, propagating, or transporting a program for use by or in connection with an instruction execution system, apparatus, or device) comprising:
performing, during a first state of a two-stage process:
retrieving…a schema of the source table from the first database and a schema of the plurality of target tables from the second database (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review, retrieval-enhanced large language model (LLM) combines the capabilities of a LLM with a retrieval mechanism to improve the relevance, coherence, and factual accuracy of generate text, producing prompts containing source schema attributes, such as columns, with relevant target schema candidates; here Mishaeli does not explicitly disclose in runtime and retrieving a set of metadata comprising a schema, but the Shah, Klein, and Dines references disclose the features, as discussed below);…
identifying…one or more matching tables among the plurality of target tables…wherein the identifying comprises constructing a first prompt based on the schemas of the source table and the schema of the plurality of target tables, and sending the first prompt to a large language model (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source database has a first schema and a target database has a second schema, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), the set of target table documents includes a top number of target attributes selected based on the prompt, a ranked list of top potential matches generated for further review, retrieval-enhanced large language model (LLM) combines the capabilities of a LLM with a retrieval mechanism to improve the relevance, coherence, and factual accuracy of generate text, producing prompts containing source schema attributes, such as columns, with relevant target schema candidates; Mishaeli in [0020] and [0038] discloses retrieving contextually relevant information from knowledge base based on input prompt and using this information to inform and guide generation of a response, a prompt created containing source attributes and target candidate attributes, which is used by LLM to select top K most similar target attributes from top documents that represent the candidate target tables; here Mishaeli does not explicitly disclose in runtime and identifying based on the filtered metadata, but the Shah and Dines references disclose the features, as discussed below);
receiving a response from the large language model comprising one or more matching tables based on first…embeddings from the schema of the source table and second…embeddings from the schema for the plurality of target tables meeting at least a threshold similarity metric (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review, retrieval-enhanced large language model (LLM) combines the capabilities of a LLM with a retrieval mechanism to improve the relevance, coherence, and factual accuracy of generate text, producing prompts containing source schema attributes, such as columns, with relevant target schema candidates; Mishaeli in [0023] and [0031] discloses representation of text via embeddings allows for efficient and accurate passage retrieval using semantic similarity, a text embedding model is utilized to encode both the candidate source attribute, accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis of measuring semantic similarity, enabling the efficient retrieval of candidate tables; here Mishaeli does not explicitly disclose vector embeddings and meeting a threshold similarity metric, but the Klein reference discloses the feature, as discussed below);
performing, during a second stage of the two-stage process:
obtaining…first sample attribute data from the source table and second sample attribute data from a selected matching table (Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.); and
identifying…one or more pairs of matching attributes between the source table and the selected matching table based on comparison of the first sample attribute data and the second sample attribute data using the large language model (Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.).
Mishaeli discloses retrieving schema, identifying matching tables, obtaining sample attributes, and identifying matching attributes, however, Mishaeli does not explicitly disclose:
retrieving…a set of metadata comprising a schema…;
filtering… the set of metadata based, at least in part, on semantic context to generate a reduced set of metadata;
identifying…based on the filtered metadata…;
The Shah reference discloses retrieving a set of metadata comprising a schema, filtering the set of metadata based, at least in part, on semantic context to generate a reduced set of metadata, and identifying based on the filtered metadata (Shah in [0050], [0055], and [0057] discloses populating a vector score from extracted metadata from data sources, the vector store providing embeddings that enable identification of data sources where relevant information should be found, vector score includes embeddings including queries and responses for each of a plurality of data sources as well as descriptions representing the schema and content of each of the data sources, a metadata store includes raw text for a schema, a prompt includes metadata of data sources representing their schema and content, prompt includes metadata representing schema and content of data source from the vector store and/or the metadata store; Shah in [0074] and [0095] discloses context retrieval identifying filtered metadata of data sets represented in the metadata store that are applicable to user’s question and matching them to validated queries in data stores, query gets semantically matched with existing sample questions and other embedded data in a vector store and/or metadata store, filtered metadata includes nodes and relationships, table and index data, or other types of data, identified filtered metadata used to generate prompt, the filtered metadata representing schema and characteristics of data within a data source, here filtered metadata represents a reduced set of metadata; Shah in [0076], [0088], and [0116] discloses metadata includes schema of a data store and relationships between nodes of the data store, metadata for a database includes index, schema, and tables, extracting schema and identifying relationships between nodes, tables, and properties, extracting schema and metadata from data source);
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teachings of Mishaeli and Shah, to have combined Mishaeli and Shah. The motivation to combine Mishaeli and Shah would be to search across multiple data sources by generating one or more custom queries targeted and formatted to one or more data sources (Shah: [0001] and [0013]).
Mishaeli discloses embeddings allowing for efficient and accurate passage retrieval using semantic similarity and Shah discloses a vector store providing embeddings that enable identification of data sources where relevant information can be found, however, Mishaeli and Shah do not explicitly disclose:
receiving a response…based on first vector embeddings…and second vector embeddings…meeting at least a threshold similarity metric;
The Klein reference discloses receiving a response based on first vector embeddings and second vector embeddings meeting at least a threshold similarity metric (Klen in [0004] and [0006] discloses determining a request embedding based on a request, the request embedding is compared to an embedding to determine a similarity satisfying a similarity criteria, descriptions of tables of a database are received, the descriptions are provided to an embedding model configured to generate embeddings based on input data; Klein in [0039] discloses embeddings being represented as vectors, the distance between two embeddings in vector space is correlated with semantic similarity between two inputs in their original format; Klein in [0041] and [0078] discloses similarity criteria specifies a threshold to be satisfied by a measure of similarity between embeddings, such as a level of similarity between the embeddings meets or exceeds a threshold).
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teachings of Mishaeli, Shah, and Klein, to have combined Mishaeli, Shah, and Klein. The motivation to combine Mishaeli, Shah, and Klein would be to simplify interaction between a user or application desiring to access or manipulate data and a database by converting a provided query to a query suitable for execution against the database (Klein: [0034]).
Mishaeli discloses retrieving schemas of source and target tables, identifying matching tables based on a constructed prompt, sending the prompt to a LLM, receiving a response, obtaining attribute data from tables, and identifying pairs of matching attributes between tables based on comparison of attribute data using the LLM, Shah discloses retrieving a set of metadata comprising a schema, filtering the set of metadata, and identifying based on the filtered metadata, Klein discloses receiving a response based on first and second vector embeddings meeting a threshold similarity metric, however, Mishaeli, Shah, and Klein do not explicitly disclose:
…in runtime…;
The Dines reference discloses in runtime (Dines in [0159] and [0160] discloses performing semantic matching between a source and a target at runtime, performing schema identification and semantic mapping at runtime; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime).
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teachings of Mishaeli, Shah, Klein, and Dines, to have combined Mishaeli, Shah, Klein, and Dines. The motivation to combine Mishaeli, Shah, Klein, and Dines would be to accurately determine semantic relationships for various targets by performing target-based schema identification and/or semantic mapping (Dines: [0003] and [0004]).
With respect to claim 2, Mishaeli in view of Shah in view of Klein and in further view of Dines discloses the computing system of claim 1, wherein identifying the one or more matching tables comprises:
constructing, in runtime, the first prompt, wherein constructing the first prompt comprises inserting the schema of the source table and the schema of the plurality of target tables into a first prompt template (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review; Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime); and
prompting, in runtime, the large language model using the first prompt (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review; Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime).
With respect to claim 3, Mishaeli in view of Shah in view of Klein and in further view of Dines discloses the computing system of claim 2, wherein constructing the first prompt further comprises removing some attributes in the schema of the source table or the schema of the plurality of target tables from the first prompt based on one or more predefined filtering criteria (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review; Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime).
With respect to claim 4, Mishaeli in view of Shah in view of Klein and in further view of Dines discloses the computing system of claim 1,
wherein retrieving the schema of the source table comprises obtaining, in runtime, text descriptions of the source table and attributes of the source table from a first dictionary associated with the first database (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review; Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime),
wherein retrieving the schema of the plurality of target tables comprises obtaining, in runtime, text descriptions of the plurality of target tables and attributes of the plurality of target tables from a second dictionary associated with the second database (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review; Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime).
With respect to claim 5, Mishaeli in view of Shah in view of Klein and in further view of Dines discloses the computing system of claim 1,
wherein retrieving the schema of the source table comprises obtaining, in runtime, statistics of the source table (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors),
wherein retrieving the schema of the plurality of target tables comprises obtaining, in runtime, statistics of the plurality of target tables (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors).
With respect to claim 6, Mishaeli in view of Shah in view of Klein and in further view of Dines discloses the computing system of claim 1, wherein identifying the one or more matching tables comprises:
generating a first vector embedding based on the schema of the source table (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors); and
generating second vector embeddings based on the schema of the plurality of target tables (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors).
With respect to claim 7, Mishaeli in view of Shah in view of Klein and in further view of Dines discloses the computing system of claim 6, wherein identifying the one or more matching tables further comprises:
measuring table similarities between the first vector embedding and the second vector embeddings (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors); and
identifying table similarities that are greater than a predefined threshold (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors).
With respect to claim 8, Mishaeli in view of Shah in view of Klein and in further view of Dines discloses the computing system of claim 1, wherein identify one or more pairs of matching attributes comprises:
constructing, in runtime, a second prompt (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review; Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime); and
prompting, in runtime, the large language model using the second prompt (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review; Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime),
wherein constructing the second prompt comprises inserting the first sample attribute data and second sample attribute data into a second prompt template (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review; Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime).
With respect to claim 9, Mishaeli in view of Shah in view of Klein and in further view of Dines discloses the computing system of claim 8, wherein constructing the second prompt further comprises removing some of the first sample attribute data or the second sample attribute data from the second prompt based on one or more predefined filtering criteria (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review; Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime).
With respect to claim 10, Mishaeli in view of Shah in view of Klein and in further view of Dines discloses the computing system of claim 1, wherein identify one or more pairs of matching attributes comprises:
generating third vector embeddings based on the first sample attribute data (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors);
generating fourth vector embeddings based on the second sample attribute data (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors); and
measuring attribute similarities between the third vector embeddings and the fourth vector embeddings (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors).
With respect to claim 11, Mishaeli discloses a computer-implemented method for improved schema matching of two databases (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema), the method comprising:
performing, during a first state of a two-stage process:
retrieving…a schema of the source table from the first database and a schema of the plurality of target tables from the second database (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review, retrieval-enhanced large language model (LLM) combines the capabilities of a LLM with a retrieval mechanism to improve the relevance, coherence, and factual accuracy of generate text, producing prompts containing source schema attributes, such as columns, with relevant target schema candidates; here Mishaeli does not explicitly disclose in runtime and retrieving a set of metadata comprising a schema, but the Shah, Klein, and Dines references disclose the features, as discussed below);…
identifying…one or more matching tables among the plurality of target tables…wherein the identifying comprises constructing a first prompt based on the schemas of the source table and the schema of the plurality of target tables, and sending the first prompt to a large language model (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source database has a first schema and a target database has a second schema, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), the set of target table documents includes a top number of target attributes selected based on the prompt, a ranked list of top potential matches generated for further review, retrieval-enhanced large language model (LLM) combines the capabilities of a LLM with a retrieval mechanism to improve the relevance, coherence, and factual accuracy of generate text, producing prompts containing source schema attributes, such as columns, with relevant target schema candidates; Mishaeli in [0020] and [0038] discloses retrieving contextually relevant information from knowledge base based on input prompt and using this information to inform and guide generation of a response, a prompt created containing source attributes and target candidate attributes, which is used by LLM to select top K most similar target attributes from top documents that represent the candidate target tables; here Mishaeli does not explicitly disclose in runtime and identifying based on the filtered metadata, but the Shah and Dines references disclose the features, as discussed below);
receiving a response from the large language model comprising one or more matching tables based on first…embeddings from the schema of the source table and second…embeddings from the schema for the plurality of target tables meeting at least a threshold similarity metric (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review, retrieval-enhanced large language model (LLM) combines the capabilities of a LLM with a retrieval mechanism to improve the relevance, coherence, and factual accuracy of generate text, producing prompts containing source schema attributes, such as columns, with relevant target schema candidates; Mishaeli in [0023] and [0031] discloses representation of text via embeddings allows for efficient and accurate passage retrieval using semantic similarity, a text embedding model is utilized to encode both the candidate source attribute, accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis of measuring semantic similarity, enabling the efficient retrieval of candidate tables; here Mishaeli does not explicitly disclose vector embeddings and meeting a threshold similarity metric, but the Klein reference discloses the feature, as discussed below);
performing, during a second stage of the two-stage process:
obtaining…first sample attribute data from the source table and second sample attribute data from a selected matching table (Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.); and
identifying…one or more pairs of matching attributes between the source table and the selected matching table based on comparison of the first sample attribute data and the second sample attribute data using the large language model (Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.).
Mishaeli discloses retrieving schema, identifying matching tables, obtaining sample attributes, and identifying matching attributes, however, Mishaeli does not explicitly disclose:
retrieving…a set of metadata comprising a schema…;
filtering… the set of metadata based, at least in part, on semantic context to generate a reduced set of metadata;
identifying…based on the filtered metadata…;
The Shah reference discloses retrieving a set of metadata comprising a schema, filtering the set of metadata based, at least in part, on semantic context to generate a reduced set of metadata, and identifying based on the filtered metadata (Shah in [0050], [0055], and [0057] discloses populating a vector score from extracted metadata from data sources, the vector store providing embeddings that enable identification of data sources where relevant information should be found, vector score includes embeddings including queries and responses for each of a plurality of data sources as well as descriptions representing the schema and content of each of the data sources, a metadata store includes raw text for a schema, a prompt includes metadata of data sources representing their schema and content, prompt includes metadata representing schema and content of data source from the vector store and/or the metadata store; Shah in [0074] and [0095] discloses context retrieval identifying filtered metadata of data sets represented in the metadata store that are applicable to user’s question and matching them to validated queries in data stores, query gets semantically matched with existing sample questions and other embedded data in a vector store and/or metadata store, filtered metadata includes nodes and relationships, table and index data, or other types of data, identified filtered metadata used to generate prompt, the filtered metadata representing schema and characteristics of data within a data source, here filtered metadata represents a reduced set of metadata; Shah in [0076], [0088], and [0116] discloses metadata includes schema of a data store and relationships between nodes of the data store, metadata for a database includes index, schema, and tables, extracting schema and identifying relationships between nodes, tables, and properties, extracting schema and metadata from data source);
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teachings of Mishaeli and Shah, to have combined Mishaeli and Shah. The motivation to combine Mishaeli and Shah would be to search across multiple data sources by generating one or more custom queries targeted and formatted to one or more data sources (Shah: [0001] and [0013]).
Mishaeli discloses embeddings allowing for efficient and accurate passage retrieval using semantic similarity and Shah discloses a vector store providing embeddings that enable identification of data sources where relevant information can be found, however, Mishaeli and Shah do not explicitly disclose:
receiving a response…based on first vector embeddings…and second vector embeddings…meeting at least a threshold similarity metric;
The Klein reference discloses receiving a response based on first vector embeddings and second vector embeddings meeting at least a threshold similarity metric (Klen in [0004] and [0006] discloses determining a request embedding based on a request, the request embedding is compared to an embedding to determine a similarity satisfying a similarity criteria, descriptions of tables of a database are received, the descriptions are provided to an embedding model configured to generate embeddings based on input data; Klein in [0039] discloses embeddings being represented as vectors, the distance between two embeddings in vector space is correlated with semantic similarity between two inputs in their original format; Klein in [0041] and [0078] discloses similarity criteria specifies a threshold to be satisfied by a measure of similarity between embeddings, such as a level of similarity between the embeddings meets or exceeds a threshold).
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teachings of Mishaeli, Shah, and Klein, to have combined Mishaeli, Shah, and Klein. The motivation to combine Mishaeli, Shah, and Klein would be to simplify interaction between a user or application desiring to access or manipulate data and a database by converting a provided query to a query suitable for execution against the database (Klein: [0034]).
Mishaeli discloses retrieving schemas of source and target tables, identifying matching tables based on a constructed prompt, sending the prompt to a LLM, receiving a response, obtaining attribute data from tables, and identifying pairs of matching attributes between tables based on comparison of attribute data using the LLM, Shah discloses retrieving a set of metadata comprising a schema, filtering the set of metadata, and identifying based on the filtered metadata, Klein discloses receiving a response based on first and second vector embeddings meeting a threshold similarity metric, however, Mishaeli, Shah, and Klein do not explicitly disclose:
…in runtime…;
The Dines reference discloses in runtime (Dines in [0159] and [0160] discloses performing semantic matching between a source and a target at runtime, performing schema identification and semantic mapping at runtime; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime).
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teachings of Mishaeli, Shah, Klein, and Dines, to have combined Mishaeli, Shah, Klein, and Dines. The motivation to combine Mishaeli, Shah, Klein, and Dines would be to accurately determine semantic relationships for various targets by performing target-based schema identification and/or semantic mapping (Dines: [0003] and [0004]).
With respect to claim 12, Mishaeli in view of Shah in view of Klein and in further view of Dines discloses the computer-implemented method of claim 11, wherein identifying the one or more matching tables comprises:
constructing, in runtime, the first prompt, wherein constructing the first prompt comprises inserting the schema of the source table and the schema of the plurality of target tables into a first prompt template (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review; Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime); and
prompting, in runtime, the large language model using the first prompt (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review; Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime).
With respect to claim 13, Mishaeli in view of Shah in view of Klein and in further view of Dines discloses the computer-implemented method of claim 11,
wherein retrieving the schema of the source table comprises obtaining, in runtime, text descriptions of the source table and attributes of the source table from a first dictionary associated with the first database (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review; Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime),
wherein retrieving the schema of the plurality of target tables comprises obtaining, in runtime, text descriptions of the plurality of target tables and attributes of the plurality of target tables from a second dictionary associated with the second database (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review; Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime).
With respect to claim 14, Mishaeli in view of Shah in view of Klein and in further view of Dines discloses the computer-implemented method of claim 11,
wherein retrieving the schema of the source table comprises obtaining, in runtime, statistics of the source table (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors),
wherein retrieving the schema of the plurality of target tables comprises obtaining, in runtime, statistics of the plurality of target tables (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors).
With respect to claim 15, Mishaeli in view of Shah in view of Klein and in further view of Dines discloses the computer-implemented method of claim 11, wherein identifying the one or more matching tables comprises:
generating a first vector embedding based on the schema of the source table (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors); and
generating second vector embeddings based on the schema of the plurality of target tables (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors).
With respect to claim 16, Mishaeli in view of Shah in view of Klein and in further view of Dines discloses the computer-implemented method of claim 15, wherein identifying the one or more matching tables further comprises:
measuring table similarities between the first vector embedding and the second vector embeddings (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors); and
ranking the one or more matching tables based on the table similarities (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors).
With respect to claim 17, Mishaeli in view of Shah in view of Klein and in further view of Dines discloses the computer-implemented method of claim 11, wherein identify one or more pairs of matching attributes comprises:
constructing, in runtime, a second prompt (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review; Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime); and
prompting, in runtime, the large language model using the second prompt (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review; Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime),
wherein constructing the second prompt comprises inserting the first sample attribute data and second sample attribute data into a second prompt template (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review; Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime).
With respect to claim 18, Mishaeli in view of Shah in view of Klein and in further view of Dines discloses the computer-implemented method of claim 11, wherein identify one or more pairs of matching attributes comprises:
generating third vector embeddings based on the first sample attribute data (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors); and
generating fourth vector embeddings based on the second sample attribute data (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors).
With respect to claim 19, Mishaeli in view of Shah in view of Klein and in further view of Dines discloses the computer-implemented method of claim 18, wherein identify one or more pairs of matching attributes further comprises:
measuring attribute similarities between the third vector embeddings and the fourth vector embeddings (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors); and
ranking the one or more pairs of matching attributes based on the attribute similarities (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, structured document containing all textual description and information of the schemas; Mishaeli in [0023] and [0031] discloses embedding models provide contextual embeddings, representation of text via embeddings allow for efficient and accurate passage retrieval using semantic similarity, to retrieve candidate tables a text embedding model encodes both the candidate source attribute accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis for measuring semantic similarity; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.; Dines in [0028] discloses measuring similarity based on a similarity threshold, a high similarity threshold indicates a requirement for an almost exact match, threshold can be re-scaled or adjusted; Dines in [0052] and [0057] discloses performing various functions such as statistical modeling, collecting statistical data; Dines in [0034] and [0059] discloses matching identical or similar names/phrases in a target to those in source data, processing tables; Dines in [0134] discloses processing column vector input; Dines in [0164] and [0165] discloses performing target based schema identification and semantic matching between a source and target at runtime, schema matching using natural language processing, converting words into vectors).
With respect to claim 20, Mishaeli discloses one or more non-transitory computer-readable media having encoded thereon computer-executable instructions causing one or more processors to perform a method for improved schema matching of two databases (Mishaeli in [0052] and [0063] discloses stored instructions executed by one or more processors and one or more memory included within computer system Mishaeli in [0059] and [0060] discloses computer program product on a computer-usable storage medium having computer-usable program code embodied in the medium, medium containing, storing, communicating, propagating, or transporting a program for use by or in connection with an instruction execution system, apparatus, or device), the method comprising:
performing, during a first state of a two-stage process:
retrieving…a schema of the source table from the first database and a schema of the plurality of target tables from the second database (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source attribute in a source table of the source schema represented as a passage based structured document and a plurality of target tables of the target schema are represented as passage based structured documents, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review, retrieval-enhanced large language model (LLM) combines the capabilities of a LLM with a retrieval mechanism to improve the relevance, coherence, and factual accuracy of generate text, producing prompts containing source schema attributes, such as columns, with relevant target schema candidates; here Mishaeli does not explicitly disclose in runtime and retrieving a set of metadata comprising a schema, but the Shah, Klein, and Dines references disclose the features, as discussed below);…
identifying…one or more matching tables among the plurality of target tables…wherein the identifying comprises constructing a first prompt based on the schemas of the source table and the schema of the plurality of target tables, and sending the first prompt to a large language model (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, a source database has a first schema and a target database has a second schema, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), the set of target table documents includes a top number of target attributes selected based on the prompt, a ranked list of top potential matches generated for further review, retrieval-enhanced large language model (LLM) combines the capabilities of a LLM with a retrieval mechanism to improve the relevance, coherence, and factual accuracy of generate text, producing prompts containing source schema attributes, such as columns, with relevant target schema candidates; Mishaeli in [0020] and [0038] discloses retrieving contextually relevant information from knowledge base based on input prompt and using this information to inform and guide generation of a response, a prompt created containing source attributes and target candidate attributes, which is used by LLM to select top K most similar target attributes from top documents that represent the candidate target tables; here Mishaeli does not explicitly disclose in runtime and identifying based on the filtered metadata, but the Shah and Dines references disclose the features, as discussed below);
receiving a response from the large language model comprising one or more matching tables based on first…embeddings from the schema of the source table and second…embeddings from the schema for the plurality of target tables meeting at least a threshold similarity metric (Mishaeli in [0019] and [0029] discloses matching a source database schema to a target database schema, plurality of target table documents are searched to retrieve a set of target table documents that include target attributes that correspond to the source attribute, a set of target attribute is selected based on a prompt using a retrieval-enhanced large language model (LLM), a ranked list of top potential matches generated for further review, retrieval-enhanced large language model (LLM) combines the capabilities of a LLM with a retrieval mechanism to improve the relevance, coherence, and factual accuracy of generate text, producing prompts containing source schema attributes, such as columns, with relevant target schema candidates; Mishaeli in [0023] and [0031] discloses representation of text via embeddings allows for efficient and accurate passage retrieval using semantic similarity, a text embedding model is utilized to encode both the candidate source attribute, accompanied by its table description, and the corpus of target table documents, these embeddings serve as a basis of measuring semantic similarity, enabling the efficient retrieval of candidate tables; here Mishaeli does not explicitly disclose vector embeddings and meeting a threshold similarity metric, but the Klein reference discloses the feature, as discussed below);
performing, during a second stage of the two-stage process:
obtaining…first sample attribute data from the source table and second sample attribute data from a selected matching table (Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.); and
identifying…one or more pairs of matching attributes between the source table and the selected matching table based on comparison of the first sample attribute data and the second sample attribute data using the large language model (Mishaeli in [0032] and [0038] discloses LLM tasked with selecting top most similar target attributes from the set of retrieved tables, model assesses similarity based on context provided by document representations, yielding ranked list of potential matches for each attribute in the source schema, a prompt is created containing source attributes and target candidate attributes which is used by LLM to select top most similar target attributes from the top documents that represent candidate target tables; Mishaeli in [0039] and [0040] discloses prompt created using artificial intelligence, prompt includes source schema information and target schema information and provides schema matching instructions, prompt includes a text description of a desired schema matching outcome, source table, source attribute, source table name, source column name, target tables, names, attributes etc.).
Mishaeli discloses retrieving schema, identifying matching tables, obtaining sample attributes, and identifying matching attributes, however, Mishaeli does not explicitly disclose:
retrieving…a set of metadata comprising a schema…;
filtering… the set of metadata based, at least in part, on semantic context to generate a reduced set of metadata;
identifying…based on the filtered metadata…;
The Shah reference discloses retrieving a set of metadata comprising a schema, filtering the set of metadata based, at least in part, on semantic context to generate a reduced set of metadata, and identifying based on the filtered metadata (Shah in [0050], [0055], and [0057] discloses populating a vector score from extracted metadata from data sources, the vector store providing embeddings that enable identification of data sources where relevant information should be found, vector score includes embeddings including queries and responses for each of a plurality of data sources as well as descriptions representing the schema and content of each of the data sources, a metadata store includes raw text for a schema, a prompt includes metadata of data sources representing their schema and content, prompt includes metadata representing schema and content of data source from the vector store and/or the metadata store; Shah in [0074] and [0095] discloses context retrieval identifying filtered metadata of data sets represented in the metadata store that are applicable to user’s question and matching them to validated queries in data stores, query gets semantically matched with existing sample questions and other embedded data in a vector store and/or metadata store, filtered metadata includes nodes and relationships, table and index data, or other types of data, identified filtered metadata used to generate prompt, the filtered metadata representing schema and characteristics of data within a data source, here filtered metadata represents a reduced set of metadata; Shah in [0076], [0088], and [0116] discloses metadata includes schema of a data store and relationships between nodes of the data store, metadata for a database includes index, schema, and tables, extracting schema and identifying relationships between nodes, tables, and properties, extracting schema and metadata from data source);
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teachings of Mishaeli and Shah, to have combined Mishaeli and Shah. The motivation to combine Mishaeli and Shah would be to search across multiple data sources by generating one or more custom queries targeted and formatted to one or more data sources (Shah: [0001] and [0013]).
Mishaeli discloses embeddings allowing for efficient and accurate passage retrieval using semantic similarity and Shah discloses a vector store providing embeddings that enable identification of data sources where relevant information can be found, however, Mishaeli and Shah do not explicitly disclose:
receiving a response…based on first vector embeddings…and second vector embeddings…meeting at least a threshold similarity metric;
The Klein reference discloses receiving a response based on first vector embeddings and second vector embeddings meeting at least a threshold similarity metric (Klen in [0004] and [0006] discloses determining a request embedding based on a request, the request embedding is compared to an embedding to determine a similarity satisfying a similarity criteria, descriptions of tables of a database are received, the descriptions are provided to an embedding model configured to generate embeddings based on input data; Klein in [0039] discloses embeddings being represented as vectors, the distance between two embeddings in vector space is correlated with semantic similarity between two inputs in their original format; Klein in [0041] and [0078] discloses similarity criteria specifies a threshold to be satisfied by a measure of similarity between embeddings, such as a level of similarity between the embeddings meets or exceeds a threshold).
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teachings of Mishaeli, Shah, and Klein, to have combined Mishaeli, Shah, and Klein. The motivation to combine Mishaeli, Shah, and Klein would be to simplify interaction between a user or application desiring to access or manipulate data and a database by converting a provided query to a query suitable for execution against the database (Klein: [0034]).
Mishaeli discloses retrieving schemas of source and target tables, identifying matching tables based on a constructed prompt, sending the prompt to a LLM, receiving a response, obtaining attribute data from tables, and identifying pairs of matching attributes between tables based on comparison of attribute data using the LLM, Shah discloses retrieving a set of metadata comprising a schema, filtering the set of metadata, and identifying based on the filtered metadata, Klein discloses receiving a response based on first and second vector embeddings meeting a threshold similarity metric, however, Mishaeli, Shah, and Klein do not explicitly disclose:
…in runtime…;
The Dines reference discloses in runtime (Dines in [0159] and [0160] discloses performing semantic matching between a source and a target at runtime, performing schema identification and semantic mapping at runtime; Dines in [0164] discloses performing target-based schema identification and semantic matching between a source and a target at runtime).
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date of the claimed invention, having the teachings of Mishaeli, Shah, Klein, and Dines, to have combined Mishaeli, Shah, Klein, and Dines. The motivation to combine Mishaeli, Shah, Klein, and Dines would be to accurately determine semantic relationships for various targets by performing target-based schema identification and/or semantic mapping (Dines: [0003] and [0004]).
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.
Remarks
The relevant prior art of record that are not used in claim rejections but are pertinent to the claims or disclosure are:
Stepin (US Pub 2026/0004602), which discloses embeddings in vector space, at runtime, prompt tuning for LLM, and use of a threshold.
Larson (US Pub 2025/0328565), which discloses embedding as a vector representation, embedding search within similarity threshold, executing LLM on a prompt for relevant textual data of a table, and runtime environments.
Ardhanari (US Pat 12,182,311) discloses table data filter including metadata enriching semantic analysis and contextual interpretation.
Contact Information
Any inquiry concerning this communication or earlier communications from the examiner should be directed to REZWANUL MAHMOOD whose telephone number is (571)272-5625. The examiner can normally be reached M-F 9-5:30.
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 J. 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.
/R.M/Examiner, Art Unit 2159
/ANN J LO/Supervisory Patent Examiner, Art Unit 2159
Prosecution Insights