Predictive injection of conversation fillers for assistant systems

1. A method comprising, by a client system:

receiving, at the client system, a first user input;

processing, by the client system, the first user input to provide an initial response by:

identifying one or more entities referenced by the first user input; and

providing, by the client system, the initial response, wherein the initial response comprises a conversational filler referencing at least one of the one or more identified entities;

processing the first user input to provide a complete response by:

identifying, by the client system, one or more intents and one or more slots associated with the first user input based on a semantic analysis by a natural-language understanding module; and

providing, by the client system, the complete response subsequent to the initial response, wherein the complete response is based on the one or more intents and the one or more slots.

2. The method of claim 1 , further comprising:

determining a classification of the first user input based on a machine-learning classifier model;

generating a plurality of candidate conversational fillers based on the classification of the first user input and the one or more identified entities, wherein each candidate conversational filler references at least one of the one or more identified entities; and

prior to providing the initial response, selecting the conversational filler from the plurality of candidate conversational fillers.

3. The method of claim 2 , wherein the candidate conversational fillers are based on conversational filler templates stored in a database of conversational filler templates, and wherein the conversational filler templates are pre-generated filler templates.

4. The method of claim 2 , wherein the candidate conversational fillers are based on conversational filler templates stored in a database of conversational filler templates, wherein the filler templates are based on content extracted from comments associated with a plurality of entities.

5. The method of claim 4 , wherein generating the plurality of candidate conversational fillers comprises selecting one or more general comments associated with the one or more identified entities from the database of general comments.

6. The method of claim 2 , wherein the classification of the first user input comprises one or more of an opinion query, a fact-based query, or a command query.

7. The method of claim 2 , wherein the classification is determined from a plurality of classifications, and wherein each classification of the plurality of classifications is associated with a set of n-grams relevant to the respective classification, wherein the first user input comprises one or more n-grams, and wherein the classification of the first user input is further based on a comparison between the one or more n-grams of the first user input to the sets of n-grams associated with the plurality of classifications.

8. The method of claim 2 , further comprising:

ranking, by the one or more computing systems, the candidate conversational fillers based on a relevancy of the candidate conversational filler to the first user input and a recency of previously used candidate conversational fillers,

wherein ranking of the candidate conversational fillers is further based on a user profile of the first user, wherein a language register associated with the first user is stored in the user profile of the first user, and wherein the ranking of the candidate conversational fillers is further based on which candidate conversational fillers match the language register associated with the first user.

9. The method of claim 2 , further comprising:

identifying a query associated with the first user input responsive to receiving the first user input; and

determining a time period to process the first user input based on the query associated with the first user input.

10. The method of claim 9 , further comprising:

ranking, by the one or more computing systems, the candidate conversational fillers based on a relevancy of the candidate conversational filler to the first user input and a recency of previously used candidate conversational fillers,

wherein ranking of each of the candidate conversational fillers is further based on a comparison between the time period to process the query and a time period to respond with the respective candidate filler.

11. The method of claim 9 , further comprising:

identifying one or more search results corresponding to the query; and

generating the complete response based on the one or more search results.

12. The method of claim 2 , further comprising:

performing, for each candidate conversational filler, a logical consistency check of the candidate conversational filler with respect to the one or more identified entities.

13. The method of claim 12 , further comprising:

ranking, by the one or more computing systems, the candidate conversational fillers based on a relevancy of the candidate conversational filler to the first user input and a recency of previously used candidate conversational fillers,

wherein ranking the candidate conversational fillers is further based on the logical consistency checks.

14. The method of claim 12 , wherein generating the plurality of candidate conversational fillers comprises filtering one or more candidate conversational fillers based on the logical consistency checks.

15. The method of claim 2 , further comprising:

ranking, by the one or more computing systems, the candidate conversational fillers based on a relevancy of the candidate conversational filler to the first user input and a recency of previously used candidate conversational fillers,

wherein information associated with the recency of previously used candidate conversational fillers provides information associated with a time period of each of the plurality of candidate conversational fillers that have been previously presented as the initial response to the first user.

16. The method of claim 15 , wherein ranking the candidate conversational fillers based on the recency of previously used candidate conversational fillers comprises filtering out one or more candidate conversational fillers that have been previously presented as the initial response to the first user within a threshold time period.

17. One or more computer-readable non-transitory storage media embodying software that is operable when executed to:

receive, at a client system, a first user input;

process, by the client system, the first user input to provide an initial response by:

identify one or more entities referenced by the first user input; and

provide, by the client system, the initial response, wherein the initial response comprises a conversational filler referencing at least one of the one or more identified entities;

process the first user input to provide a complete response by:

identify, by the client system, one or more intents and one or more slots associated with the first user input based on a semantic analysis by a natural-language understanding module; and

provide, by the client system, the complete response subsequent to the initial response, wherein the complete response is based on the one or more intents and the one or more slots.

18. A system comprising: one or more processors; and a non-transitory memory coupled to the processors comprising instructions executable by the processors, the processors operable when executing the instructions to:

receive, at a client system, a first user input;

process, by the client system, the first user input to provide an initial response by:

identify one or more entities referenced by the first user input; and

provide, by the client system, the initial response, wherein the initial response comprises a conversational filler referencing at least one of the one or more identified entities;

process the first user input to provide a complete response by:

identify, by the client system, one or more intents and one or more slots associated with the first user input based on a semantic analysis by a natural-language understanding module; and

provide, by the client system, the complete response subsequent to the initial response, wherein the complete response is based on the one or more intents and the one or more slots.