Automatically generating personalized and context-aware explanation formats for a recommendation
Embodiments of the invention are directed to a computer-implemented method. A non-limiting example of the computer-implemented method includes accessing, using an explanation generator module of a processor system, information of a recommendation associated with an application, information of the application, and information of a user of the application. The explanation generator module of the processor system is used to determine an explanation format of an explanation of the recommendation based at least in part on the information of the recommendation associated with the application, the information of the application, and the information of the user of the application.
1. A computer-implemented method comprising:
executing, using an interaction manager module of a recommender system, user feedback operations comprising receiving and storing user feedback on explanation formats of recommendation-explanation pairs generated by a recommender of the recommender system;
wherein the recommendation-explanation pairs comprise a first recommendation-explanation pair; and
determining, using an explanation generator module of the recommender system, a predicted explanation format of the first recommendation-explanation pair;
wherein the explanation generator module comprises a machine learning model trained based at least in part on the user feedback on the explanation formats of the recommendation-explanation pairs generated by the recommender.
2. The computer-implemented method of claim 1 , wherein the predicted explanation format of the first recommendation-explanation pair is based at least in part on a prediction, using the machine learning model, that a user of the recommender system will accept the predicted explanation format.
3. The computer-implemented method of claim 1 , wherein the predicted explanation format of the first recommendation-explanation pair is based at least in part on a prediction that a user of the recommender system will act on a recommendation portion of the first recommendation-explanation pair.
4. The computer-implemented method of claim 1 further comprising using the predicted explanation format type to generate a personalized combination output associated with the first recommendation-explanation pair, wherein the personalized combination output comprises an explanation format and an explanation component.
5. The computer-implemented method of claim 1 , wherein:
the recommendation-explanation pairs are associated with one or more applications; and
the machine learning model is further trained based at least in part on context information comprising:
information of a plurality of recommendations associated with the one or more applications;
information of the one or more applications; and
information of a user of the recommendation system.
6. The computer-implemented method of claim 4 , wherein the explanation component comprises one or more of natural language text, a diagram, a chart, a table, video, and audio.
7. The computer-implemented method of claim 1 , wherein the interaction manager module comprises a conversational agent configured to actively solicit the user feedback from a user of the recommender system by executing a conversational loop between the conversational agent and a user of the recommendation system.
8. A computer system comprising a processor system communicatively coupled to a memory, wherein the processor system is operable to perform processor system operations comprising:
executing, using an interaction manager module of a recommender system, user feedback operations comprising receiving and storing user feedback on explanation formats of recommendation-explanation pairs generated by a recommender of the recommender system;
wherein the recommendation-explanation pairs comprise a first recommendation-explanation pair; and
determining, using an explanation generator module of the recommender system, a predicted explanation format of the first recommendation-explanation pair;
wherein the explanation generator module comprises a machine learning model trained based at least in part on the user feedback on the explanation formats of the recommendation-explanation pairs generated by the recommender.
9. The computer system of claim 8 , wherein the predicted explanation format of the first recommendation-explanation pair is based at least in part on a prediction, using the machine learning model, that a user of the recommender system will accept the predicted explanation format.
10. The computer system of claim 8 , wherein the predicted explanation format of the first recommendation-explanation pair is based at least in part on a prediction that a user of the recommender system will act on a recommendation portion of the first recommendation-explanation pair.
11. The computer system of claim 8 , wherein:
the processor system operations further comprise using the predicted explanation format type to generate a personalized combination output associated with the first recommendation-explanation pair; and
the personalized combination output comprises an explanation format and an explanation component.
12. The computer system of claim 8 , wherein;
the recommendation-explanation pairs are associated with one or more applications; and
the machine learning model is further trained based at least in part on context information comprising:
information of a plurality of recommendations associated with the one or more applications;
information of the one or more applications; and
information of a user of the recommendation system.
13. The computer system of claim 11 , wherein the explanation component comprises one or more of natural language text, a diagram, a chart, a table, video, and audio.
14. The computer system of claim 13 , wherein the interaction manager module comprises a conversational agent configured to actively solicit the user feedback from a user of the recommender system by executing a conversational loop between the conversational agent and a user of the recommendation system.
15. A computer program product comprising a computer readable program stored on a computer readable storage medium, wherein the computer readable program, when executed on a processor system, causes the processor system to perform processor system operations comprising:
executing, using an interaction manager module of a recommender system, user feedback operations comprising receiving and storing user feedback on explanation formats of recommendation-explanation pairs generated by a recommender of the recommender system;
wherein the recommendation-explanation pairs comprise a first recommendation-explanation pair; and
determining, using an explanation generator module of the recommender system, a predicted explanation format of the first recommendation-explanation pair;
wherein the explanation generator module comprises a machine learning model trained based at least in part on the user feedback on the explanation formats of the recommendation-explanation pairs generated by the recommender.
16. The computer program product of claim 15 , wherein the predicted explanation format of the first recommendation-explanation pair is based at least in part on a prediction, using the machine learning model, that a user of the recommender system will accept the predicted explanation format.
17. The computer program product of claim 15 , wherein the interaction manager module comprises a conversational agent configured to actively solicit the user feedback from a user of the recommender system by executing a conversational loop between the conversational agent and a user of the recommendation system.
18. The computer program product of claim 15 , wherein:
the processor system operations further comprise using the predicted explanation format type to generate a personalized combination output associated with the first recommendation-explanation pair; and
the personalized combination output comprises an explanation format and an explanation component.
19. The computer program product of claim 15 , wherein:
the recommendation-explanation pairs are associated with one or more applications; and
the machine learning model is further trained based at least in part on context information comprising:
information of a plurality of recommendations associated with the one or more applications;
information of the one or more applications; and
the information of a user of the recommendation system.
20. The computer program product of claim 18 , wherein the explanation component comprises one or more of natural language text, a diagram, a chart, a table, video, and audio.