Machine learning based question and answer (Q and A) assistant

1 . One or more non-transitory, computer-readable storage media comprising instructions recorded thereon, wherein the instructions, when executed by at least one data processor of a computing system, cause the computing system to:

capture, via an input control of a graphical user interface (GUI) associated with a question and answer (Q&A) assistant engine for a multimodal content management system having a block-based data structure, a natural language prompt;

tokenize the natural language prompt to generate a set of data source tokens and a set of parameter tokens,

wherein a parameter token in the set of parameter tokens is indicative of a block property in the block-based data structure;

using the set of data source tokens and the set of parameter tokens, execute a trained neural network to generate a query structured to operate on the block-based data structure;

cause the generated query to be executed against the block-based data structure to generate a result set; and

using the result set, generate and display, via the GUI, a visualization comprising a result set item, by:

generating a navigable citation relating to a result set item in the result set;

generating a narrative comprising the navigable citation; and

while displaying the input control to accept a second natural language prompt, and responsive to detecting a user interaction with the navigable citation, causing the GUI to display content of a block in the block-based data structure, wherein the block corresponds to the result set item.

2 . The media of claim 1 , wherein the instructions, when executed by the at least one data processor of the computing system, cause the computing system to generate the query structured to operate on the block-based data structure of the multimodal content management system by performing operations to:

generate a first vectorized item relating to a block content, a block identifier, or a block property of a particular stored item in the block-based data structure;

generate a second vectorized item relating to at least one of a data source token in the set of data source tokens or the parameter token;

based on a determination that the first vectorized item and the second vectorized item meet a similarity threshold, generate a query segment that references the block content, the block identifier, or the block property of the particular stored item in the block-based data structure; and

include the query segment in the query.

3 . The media of claim 1 , wherein a parameter token in the set of parameter tokens generated based on the natural language prompt is indicative of a block property that comprises an artificial intelligence (AI)-generated item.

4 . The media of claim 1 , wherein a parameter token in the set of parameter tokens generated based on the natural language prompt is indicative of a block property that comprises at least one of: a tag relating to block content, a date, a time stamp, text, numerical information, an email address, a phone number, a formula, a roll-up, permission information, a user identifier, a file, a media item, a URL, or a link to an external source of content to query.

5 . The media of claim 1 , wherein the navigable citation is a first navigable citation, the result set item is a first result set item, and the user interaction is a first user interaction, and wherein the instructions, when executed by the at least one data processor of the computing system, cause the computing system to generate the visualization by performing operations to:

generate a second navigable citation relating to a second result set item; include the second navigable citation in the narrative; and

responsive to detecting a second user interaction with the second navigable citation, cause the GUI to display second content of a second corresponding block in the block-based data structure.

6 . The media of claim 5 , wherein the first result set item is in a first content modality and the second result set item is in a second content modality different from the first content modality, and wherein the first content modality or the second content modality comprise one or more of text, data, a table, an image, an audio item, a video item, a multimedia item, and a digital map.

7 . The media of claim 1 , wherein the instructions, when executed by the at least one data processor of the computing system, cause the computing system to generate and include in the visualization a synthetic item based on an item in the result set.

8 . A computer-implemented method, the method comprising:

capturing, via an input control of a graphical user interface (GUI) associated with a question and answer (Q&A) assistant engine for a multimodal content management system having a block-based data structure, a natural language prompt;

tokenizing the natural language prompt to generate a set of data source tokens and a set of parameter tokens,

wherein a parameter token in the set of parameter tokens is indicative of a block property in the block-based data structure;

using the set of data source tokens and the set of parameter tokens, executing a trained neural network to generate a query structured to operate on the block-based data structure;

causing the generated query to be executed against the block-based data structure to generate a result set; and

using the result set, generating and displaying, via the GUI, a visualization comprising a result set item, by:

generating a navigable citation relating to a result set item in the result set;

generating a narrative comprising the navigable citation; and

while displaying the input control to accept a second natural language prompt, and responsive to detecting a user interaction with the navigable citation, causing the GUI to display content of a block in the block-based data structure, wherein the block corresponds to the result set item.

9 . The method of claim 8 , further comprising:

generating a first vectorized item relating to a block content, a block identifier, or a block property of a particular stored item in the block-based data structure;

generating a second vectorized item relating to at least one of a data source token in the set of data source tokens or the parameter token;

based on a determination that the first vectorized item and the second vectorized item meet a similarity threshold, generating a query segment that references the block content, the block identifier, or the block property of the particular stored item in the block-based data structure; and

including the query segment in the query.

10 . The method of claim 8 , wherein a parameter token in the set of parameter tokens generated based on the natural language prompt is indicative of a block property that comprises an artificial intelligence (AI)-generated item.

11 . The method of claim 8 , wherein a parameter token in the set of parameter tokens generated based on the natural language prompt is indicative of a block property that comprises at least one of: a tag relating to block content, a date, a time stamp, text, numerical information, an email address, a phone number, a formula, a roll-up, permission information, a user identifier, a file, a media item, a URL, or an external source of content to query.

12 . The method of claim 8 , wherein the navigable citation is a first navigable citation, the result set item is a first result set item, and the user interaction is a first user interaction, the method further comprising:

generating a second navigable citation relating to a second result set item;

including the second navigable citation in the narrative; and

responsive to detecting a second user interaction with the second navigable citation, causing the GUI to display second content of a second corresponding block in the block-based data structure.

13 . The method of claim 12 , wherein the first result set item is in a first content modality and the second result set item is in a second content modality different from the first content modality, and wherein the first content modality or the second content modality comprise one or more of text, data, a table, an image, an audio item, a video item, a multimedia item, and a digital map.

14 . The method of claim 8 , further comprising generating and including in the visualization a synthetic item based on an item in the result set.

15 . A computing system comprising at least one data processor and one or more non-transitory, computer-readable storage media comprising instructions recorded thereon, wherein the instructions, when executed by the at least one data processor, cause the computing system to:

capture, via an input control of a graphical user interface (GUI) associated with a question and answer (Q&A) assistant engine for a multimodal content management system having a block-based data structure, a natural language prompt;

tokenize the natural language prompt to generate a set of data source tokens and a set of parameter tokens,

wherein a parameter token in the set of parameter tokens is indicative of a block property in the block-based data structure;

using the set of data source tokens and the set of parameter tokens, execute a trained neural network to generate a query structured to operate on the block-based data structure;

cause the generated query to be executed against the block-based data structure to generate a result set; and

using the result set, generate and display, via the GUI, a visualization comprising a result set item, by:

generating a navigable citation relating to a result set item in the result set;

generating a narrative comprising the navigable citation; and

while displaying the input control to accept a second natural language prompt, and responsive to detecting a user interaction with the navigable citation, causing the GUI to display content of a block in the block-based data structure, wherein the block corresponds to the result set item.

16 . The computing system of claim 15 , wherein the instructions, when executed by the at least one data processor of the computing system, cause the computing system to generate the query structured to operate on the block-based data structure of the multimodal content management system by performing operations to:

generate a first vectorized item relating to a block content, a block identifier, or a block property of a particular stored item in the block-based data structure;

generate a second vectorized item relating to at least one of a data source token in the set of data source tokens or the parameter token;

based on a determination that the first vectorized item and the second vectorized item meet a similarity threshold, generate a query segment that references the block content, the block identifier, or the block property of the particular stored item in the block-based data structure; and

include the query segment in the query.

17 . The computing system of claim 15 , wherein a parameter token in the set of parameter tokens generated based on the natural language prompt is indicative of a block property that comprises an artificial intelligence (AI)-generated item.

18 . The computing system of claim 15 , wherein a parameter token in the set of parameter tokens generated based on the natural language prompt is indicative of a block property that comprises at least one of: a tag relating to block content, a date, a time stamp, text, numerical information, an email address, a phone number, a formula, a roll-up, permission information, a user identifier, a file, a media item, a URL, or an external source of content to query.

19 . The computing system of claim 15 , wherein the navigable citation is a first navigable citation, the result set item is a first result set item, and the user interaction is a first user interaction, and wherein the instructions, when executed by the at least one data processor of the computing system, cause the computing system to generate the visualization by performing operations to:

generate a second navigable citation relating to a second result set item;

include the second navigable citation in the narrative; and

responsive to detecting a second user interaction with the second navigable citation, cause the GUI to display second content of a second corresponding block in the block-based data structure.

20 . The computing system of claim 19 , wherein the first result set item is in a first content modality and the second result set item is in a second content modality different from the first content modality, and wherein the first content modality or the second content modality comprise one or more of text, data, a table, an image, an audio item, a video item, a multimedia item, and a digital map.