Expandable dialogue system
A system that allows non-engineers administrators, without programming, machine language, or artificial intelligence system knowledge, to expand the capabilities of a dialogue system. The dialogue system may have a knowledge system, user interface, and learning model. A user interface allows non-engineers to utilize the knowledge system, defined by a small set of primitives and a simple language, to annotate a user utterance. The annotation may include selecting actions to take based on the utterance and subsequent actions and configuring associations. A dialogue state is continuously updated and provided to the user as the actions and associations take place. Rules are generated based on the actions, associations and dialogue state that allows for computing a wide range of results.
1. A method for training a dialogue learning model, comprising:
automatically presenting, via a user interface of a computing device, an utterance and a list of actions based on the utterance;
receiving, via the user interface of the computing device, a selection of an action from the list of actions;
receiving, via the user interface of the computing device, a designated span of the utterance less than an entirety of the utterance; and
automatically training the dialogue learning model with the action and the designated span of the utterance.
2. The method of claim 1 , wherein the utterance is a computer-generated agent utterance.
3. The method of claim 1 , wherein the utterance is a user utterance from a natural language conversation between a user and an automated assistant.
4. The method of claim 1 , wherein the utterance is a user utterance received via the user interface.
5. The method of claim 1 , wherein the action is one of a plurality of actions for responding to the utterance, and the designated span of the utterance is one of a plurality of designated spans of the utterance each corresponding to an action of the plurality of actions.
6. The method of claim 1 , wherein the automatic training configures the dialogue learning model to recognize an association between the action and natural language content in the designated span of the utterance.
7. The method of claim 6 , wherein the natural language content includes one or more tokens within the utterance.
8. The method of claim 6 , further comprising automatically generating one or more rules based on the association between the action and the natural language content.
9. The method of claim 8 , wherein the one or more rules include a rule for recognizing whether the action is applicable based on the natural language content.
10. The method of claim 8 , wherein the one or more rules include a rule for determining parameters of the action based on the natural language content.
11. The method of claim 8 , wherein the one or more rules include a rule for generating an agent utterance describing the action.
12. The method of claim 8 , wherein the dialogue learning model is a grammar induction learning model and the one or more rules are grammar rules derived based on the association by the grammar induction learning model.
13. The method of claim 8 , wherein each of the one or more rules is associated with a set of features, and the automatic training of the dialogue model include assessing relevance of the one or more features in the set of features, based on the association.
14. A method for training a dialogue learning model, comprising:
automatically presenting, via a user interface of a computing device, an utterance and a list of actions based on the utterance;
receiving, via the user interface of the computing device, a selection of a plurality of actions from the list of actions, and for each selected action, a designated span of the utterance less than an entirety of the utterance; and
automatically training the dialogue learning model with an exemplary response to the utterance including the plurality of selected actions and corresponding designated spans of the utterance.
15. The method of claim 14 , wherein the automatic training configures the dialogue learning model to recognize an association between actions and natural language content in utterances, based on the plurality of selected actions and natural language content in the corresponding designated spans of the utterance.
16. A computer system, comprising:
a processor; and
a storage device holding instructions executable by the processor to:
automatically present, via an interface of a computing device, an utterance and a list of actions based on the utterance;
receive, via the interface of the computing device, a selection of an action from the list of actions;
receive, via the interface of the computing device, a designated span of the utterance less than an entirety of the utterance; and
automatically train the dialogue learning model with the action and the designated span of the utterance.
17. The computer system of claim 16 , wherein the utterance is a computer-generated agent utterance.
18. The computer system of claim 16 , wherein the utterance is a user utterance from a natural language conversation between a user and an automated assistant.
19. The computer system of claim 16 , wherein the utterance is a user utterance received via the interface.
20. The computer system of claim 16 , wherein the action is one of a plurality of actions for responding to the utterance, and the designated span of the utterance is one of a plurality of designated spans of the utterance each corresponding to an action of the plurality of actions.