Systems and methods for enabling conversational interaction with data tables

1 . A method implemented by a virtual assistant server, the method comprising:

receiving from a user device, a natural language query targeted at a data table hosted at a user repository as part of an interaction session between the user device and the virtual assistant server;

determining one or more table intents or one or more table sub-intents from the natural language query, wherein the table intents or the table sub-intents of the natural language query are determined by providing to a natural language processing (NLP) model an intent determination prompt comprising: metadata of the data table; the natural language query; one or more instructions to the NLP model to determine the table intents or the table sub-intents; and one or more examples disclosing how to determine the table intents or the table sub-intents;

determining one or more table operations to execute on the data table based on the determined table intents or table sub-intents, wherein the one or more table operations are determined by providing one or more follow-up prompts to the NLP model, and wherein the follow-up prompts corresponding to the table intents are prioritized over the follow-up prompts corresponding to the table sub-intents while providing the follow-up prompts to the NLP model;

iteratively executing each of the determined table operations on the data table, wherein a final version of the data table results after completion of one or more of the iteratively executed determined table operations;

determining a response to the natural language query based on the final version of the data table; and

outputting the determined response to the natural language query to the user device.

2 . The method of claim 1 , further comprising:

parsing the data table;

determining metadata of the data table comprising: column names, column identifiers, and column types;

generating a structured data representation of the data table; and

storing the metadata and the structured data representation of the data table in the user repository.

3 . The method of claim 1 , wherein each of the one or more follow-up prompts to the NLP model comprises:

the natural language query;

metadata of the data table;

one or more examples disclosing how to determine the table operation for the table intent or the table sub-intent to which the follow-up corresponds to; and

one or more instructions to the NLP model to determine the table operation for the table intent or the table sub-intent to which the follow-up corresponds to.

4 . The method of claim 1 , wherein the natural language query is a selection received from the user device from one or more query suggestions, and wherein the one or more query suggestions are provided by the virtual assistant server to the user device upon receiving an indication from the user device to conversationally interact with the data table.

5 . The method of claim 4 , wherein the one or more query suggestions are received from a natural language processing (NLP) model based on a suggestion prompt provided to the NLP model by the virtual assistant server, and wherein the suggestion prompt comprises: metadata of the data table; and one or more instructions to the NLP model to generate the one or more query suggestions.

6 . The method of claim 1 , wherein each of the determined table operations is executed on a structured data representation of the data table which is iteratively updated with one or more of the executions.

7 . A virtual assistant server comprising:

one or more processors; and

a memory coupled to the one or more processors which are configured to execute programmed instructions stored in the memory to:

receive from a user device, a natural language query targeted at a data table hosted at a user repository as part of an interaction session between the user device and the virtual assistant server;

determine one or more table intents or one or more table sub-intents from the natural language query, wherein the table intents or the table sub-intents of the natural language query are determined by providing to a natural language processing (NLP) model an intent determination prompt comprising: metadata of the data table; the natural language query; one or more instructions to the NLP model to determine the table intents or the table sub-intents; and one or more examples disclosing how to determine the table intents or the table sub-intents;

determine one or more table operations to execute on the data table based on the determined table intents or table sub-intents, wherein the one or more table operations are determined by providing one or more follow-up prompts to the NLP model, and wherein the follow-up prompts corresponding to the table intents are prioritized over the follow-up prompts corresponding to the table sub-intents while providing the follow-up prompts to the NLP model;

iteratively execute each of the determined table operations on the data table, wherein a final version of the data table results after completion of one or more of the iteratively executed determined table operations;

determine a response to the natural language query based on the final version of the data table; and

output the determined response to the natural language query to the user device.

8 . The virtual assistant server of claim 7 , the one or more processors are further configured to:

parse the data table;

determine metadata of the data table comprising: column names, column identifiers, and column types;

generate a structured data representation of the data table; and

store the metadata and the structured data representation of the data table in the user repository.

9 . The virtual assistant server of claim 7 , wherein each of the one or more follow-up prompts to the NLP model comprises:

the natural language query;

metadata of the data table;

one or more examples disclosing how to determine the table operation for the table intent or the table sub-intent to which the follow-up corresponds to; and

one or more instructions to the NLP model to determine the table operation for the table intent or the table sub-intent to which the follow-up corresponds to.

10 . The virtual assistant server of claim 7 , wherein the natural language query is a selection received from the user device from one or more query suggestions, and wherein the one or more query suggestions are provided by the virtual assistant server to the user device upon receiving an indication from the user device to conversationally interact with the data table.

11 . The virtual assistant server of claim 10 , wherein the one or more query suggestions are received from a natural language processing (NLP) model based on a suggestion prompt provided to the NLP model by the virtual assistant server, and wherein the suggestion prompt comprises: metadata of the data table; and one or more instructions to the NLP model to generate the one or more query suggestions.

12 . The virtual assistant server of claim 7 , wherein each of the determined table operations is executed on a structured data representation of the data table which is iteratively updated with one or more of the executions.

13 . A non-transitory computer-readable medium storing instructions which when executed by one or more processors, causes the one or more processors to:

receive from a user device, a natural language query targeted at a data table hosted at a user repository as part of an interaction session between the user device and the virtual assistant server;

determine one or more table intents or one or more table sub-intents from the natural language query, wherein the table intents or the table sub-intents of the natural language query are determined by providing to a natural language processing (NLP) model an intent determination prompt comprising: metadata of the data table; the natural language query; one or more instructions to the NLP model to determine the table intents or the table sub-intents; and one or more examples disclosing how to determine the table intents or the table sub-intents;

determine one or more table operations to execute on the data table based on the determined table intents or table sub-intents, wherein the one or more table operations are determined by providing one or more follow-up prompts to the NLP model, and wherein the follow-up prompts corresponding to the table intents are prioritized over the follow-up prompts corresponding to the table sub-intents while providing the follow-up prompts to the NLP model;

iteratively execute each of the determined table operations on the data table, wherein a final version of the data table results after completion of one or more of the iteratively executed determined table operations;

determine a response to the natural language query based on the final version of the data table; and

output the determined response to the natural language query to the user device.

14 . The non-transitory computer-readable medium of claim 13 , further comprising instructions which when executed by the one or more processors, causes the one or more processors to:

parse the data table;

determine metadata of the data table comprising: column names, column identifiers, and column types;

generate a structured data representation of the data table; and

store the metadata and the structured data representation of the data table in the user repository.

15 . The non-transitory computer-readable medium of claim 13 , wherein each of the one or more follow-up prompts to the NLP model comprises:

the natural language query;

metadata of the data table;

one or more examples disclosing how to determine the table operation for the table intent or the table sub-intent to which the follow-up corresponds to; and

one or more instructions to the NLP model to determine the table operation for the table intent or the table sub-intent to which the follow-up corresponds to.

16 . The non-transitory computer-readable medium of claim 13 , wherein the natural language query is a selection received from the user device from one or more query suggestions, and wherein the one or more query suggestions are provided by the virtual assistant server to the user device upon receiving an indication from the user device to conversationally interact with the data table.

17 . The non-transitory computer-readable medium of claim 16 , wherein the one or more query suggestions are received from a natural language processing (NLP) model based on a suggestion prompt provided to the NLP model by the virtual assistant server, and wherein the suggestion prompt comprises: metadata of the data table; and one or more instructions to the NLP model to generate the one or more query suggestions.

18 . The non-transitory computer-readable medium of claim 13 , wherein each of the determined table operations is executed on a structured data representation of the data table which is iteratively updated with one or more of the executions.