Techniques to respond to user requests using natural-language machine learning based on branching example conversations

1. A computer-implemented method, comprising:

receiving an example-interaction repository comprising one or more pregenerated branching example user-to-bot interactions;

generating a plurality of linearized example user-to-bot interactions by traversing each of the one or more branching example user-to-bot interactions multiple times, wherein generating the plurality of linearized example user-to-bot interactions comprises:

traversing a branching example user-to-bot interaction by randomly selecting an outcome of each of a plurality of branches for the branching example user-to-bot interaction to generate a linearized example user-to-bot interaction based on the branching example user-to-bot interaction;

detecting that the linearized example user-to-bot interaction has exceeded a predefined stopping length;

stopping the traversing of the branching example user-to-bot interaction; and

adding the linearized example user-to-bot interaction to the plurality of linearized example user-to-bot interactions;

generating a linearized example-interaction repository by replacing the one or more branching example user-to-bot interactions with the linearized example user-to-bot interactions;

submitting the linearized example-interaction repository as training data to the natural-language machine learning component;

receiving a sequence model from the natural-language machine learning component in response to submitting the linearized example-interaction repository; and

performing a user-to-bot conversation based on the sequence model.

2. The method of claim 1 , wherein a branching example user-to-bot interaction comprises a branching jump instruction from a later part of the branching example user-to-bot interaction to an earlier part of the branching example user-to-bot interaction, the branching jump instruction empowering conditional repetition of a portion of the branching example user-to-bot interaction.

3. The method of claim 1 , wherein a branching example user-to-bot interaction comprises a multiway branching instruction between two or more paths forward through the branching example user-to-bot interaction, the multiway branching instruction empowering selection of a which portion of the branching example user-to-bot interaction is performed.

4. The method of claim 1 , wherein one or more of the plurality of example user-to-bot interactions includes one or more quick-reply prompts to a bot message, wherein each of the one or more quick-reply prompts is a suggested user response to the bot message.

5. The method of claim 4 , wherein one or more of the one or more quick-reply prompts is generated from production-performed user-to-bot conversations.

6. The method of claim 4 , wherein one or more of the one or more quick-reply prompts is associated with one or more branches for a branching example user-to-bot interaction.

7. An apparatus, comprising:

a processor circuit;

an interaction processing component operative on the processor circuit to generate a plurality of linearized example user-to-bot interactions by traversing each of one or more pregenerated branching example user-to-bot interactions, multiple times, wherein generating the plurality of linearized example user-to-bot interactions comprises:

traversing a branching example user-to-bot interaction by randomly selecting an outcome of each of a plurality of branches for the branching example user-to-bot interaction to generate a linearized example user-to-bot interaction based on the branching example user-to-bot interaction;

detecting that the linearized example user-to-bot interaction has exceeded a predefined stopping length;

stopping the traversing of the branching example user-to-bot interaction; and

adding the linearized example user-to-bot interaction to the plurality of linearized example user-to-bot interactions;

generate a linearized example-interaction repository by replacing the one or more branching example user-to-bot interactions with the plurality of linearized example user-to-bot interactions; and

submit the linearized example-interaction repository as training data to the natural-language machine learning component, and receive a sequence model from the natural-language machine learning component in response to submitting the linearized example-interaction repository;

a client communication component operative to perform a user message exchange with a user client to perform a user-to-bot conversation based on the sequence model; and

a bot application interface component operative to receive the example-interaction repository from a bot application and perform a bot information exchange with the bot application to perform the user-to-bot conversation based on the sequence model, wherein performing the user-to-bot conversation based on the sequence model comprises applying the sequence model to a user-request context, a user-to-bot conversation history, semantic analysis information for one or more user messages, and extracted unsupervised information for the one or more user messages.

8. The apparatus of claim 7 , wherein a branching example user-to-bot interaction comprises a branching jump instruction from a later part of the branching example user-to-bot interaction to an earlier part of the branching example user-to-bot interaction, the branching jump instruction empowering conditional repetition of a portion of the branching example user-to-bot interaction.

9. The apparatus of claim 7 , wherein a branching example user-to-bot interaction comprises a multiway branching instruction between two or more paths forward through the branching example user-to-bot interaction, the multiway branching instruction empowering selection of a which portion of the branching example user-to-bot interaction is performed.

10. The apparatus of claim 7 , wherein one or more of the plurality of example user-to-bot interactions includes one or more quick-reply prompts to a bot message, wherein each of the one or more quick-reply prompts is a suggested user response to the bot message.

11. The apparatus of claim 10 , wherein one or more of the one or more quick-reply prompts is generated from production-performed user-to-bot conversations.

12. The apparatus of claim 10 , wherein one or more of the one or more quick-reply prompts is associated with one or more branches for a branching example user-to-bot interaction.

13. At least one non-transitory computer-readable storage medium comprising instructions that, when executed, cause a system to:

receive an example-interaction repository comprising one or more pregenerated branching example user-to-bot interactions;

generate a plurality of linearized example user-to-bot interactions by traversing each of the one or more branching example user-to-bot interactions multiple times, wherein generating the plurality of linearized example user-to-bot interactions comprises:

traversing a branching example user-to-bot interaction by randomly selecting an outcome of each of a plurality of branches for the branching example user-to-bot interaction to generate a linearized example user-to-bot interaction based on the branching example user-to-bot interaction;

detecting that the linearized example user-to-bot interaction has exceeded a predefined stopping length;

stopping the traversing of the branching example user-to-bot interaction; and

adding the linearized example user-to-bot interaction to the plurality of linearized example user-to-bot interactions;

generate a linearized example-interaction repository by replacing the one or more branching example user-to-bot interactions with the linearized example user-to-bot interactions;

submit the linearized example-interaction repository as training data to a natural-language machine learning component;

receive a sequence model from the natural-language machine learning component in response to submitting the linearized example-interaction repository; and

perform a user-to-bot conversation based on the sequence model.

14. The computer-readable storage medium of claim 13 , wherein a branching example user-to-bot interaction comprises a branching jump instruction from a later part of the branching example user-to-bot interaction to an earlier part of the branching example user-to-bot interaction, the branching jump instruction empowering conditional repetition of a portion of the branching example user-to-bot interaction.

15. The computer-readable storage medium of claim 13 , wherein a branching example user-to-bot interaction comprises a multiway branching instruction between two or more paths forward through the branching example user-to-bot interaction, the multiway branching instruction empowering selection of a which portion of the branching example user-to-bot interaction is performed.

16. The computer-readable storage medium of claim 13 , wherein one or more of the plurality of example user-to-bot interactions includes one or more quick-reply prompts to a bot message, wherein each of the one or more quick-reply prompts is a suggested user response to the bot message.

17. The computer-readable storage medium of claim 16 , wherein one or more of the one or more quick-reply prompts is generated from production-performed user-to-bot conversations.

18. The computer-readable storage medium of claim 16 , wherein one or more of the one or more quick-reply prompts is associated with one or more branches for a branching example user-to-bot interaction.