US 12,608,582 B2Granted

System and method for generating conversation between artificial intelligences

AssigneeAcryl Inc.

InventorsInsik Jung, Jisung Park, Hyunho Lee, Wei Jin Park

Life of this patent

Estimating the term…

Loading patent…

Abstract

A system of generating a conversation between artificial intelligences is proposed. The system includes: first artificial intelligence including a first response generation module configured to generate a response to a start word presented by a user at a beginning of the conversation, the first artificial intelligence configured to evaluate empathy and diversity-emotion weight each for the response generated by a second response generation module of second artificial intelligence and configured to feed back an empathy evaluation score and a diversity-empathy evaluation score; and the second artificial intelligence including a second response generation module configured to generate a response to the conversation generated by the first response generation module of the first artificial intelligence, the second artificial intelligence configured to evaluate the empathy and diversity-emotion weight for the conversation generated by the first response generation module and configured to feed back the empathy evaluation score and the diversity-empathy evaluation score.

Claims10

Dependent claims

A system for generating an improved dialogue between artificial intelligences, the system comprising:

one or more processors;
and
a tangible, non-transitory memory coupled to the one or more processors and storing executable instructions that, when executed by the one or more processors, cause the system to:
receive, by a first artificial intelligence, an initial input from a user;
generate, by the first artificial intelligence, a first response to the initial input using a first response generation module;
provide, by the first artificial intelligence, the first response to a second artificial intelligence;
determine, by the second artificial intelligence, a first empathy evaluation score for the first response using a first deep learning model trained with first data identifying empathy to be learned and empathy not to be learned;
determine, by the second artificial intelligence, a first diversity-empathy evaluation score for the first response using a second deep learning model trained with second data identifying whether to generate an emotional conversation or a diverse conversation;
update, by the second artificial intelligence, a second response generation module based on the first empathy evaluation score and the first diversity-empathy evaluation score;
generate, by the second artificial intelligence, a second response to the first response using the updated second response generation module based on the first empathy evaluation score, the first diversity-empathy evaluation score, and the first response;
provide, by the second artificial intelligence, the second response to the first artificial intelligence;
determine, by the first artificial intelligence, a second empathy evaluation score for the second response using a third deep learning model trained with third data identifying empathy to be learned and empathy not to be learned;
determine, by the first artificial intelligence, a second diversity-empathy evaluation score for the second response using a fourth deep learning model trained with fourth data identifying whether to generate an emotional conversation or a diverse conversation;
update, by the first artificial intelligence, the first response generation module based on the second empathy evaluation score and the second diversity-empathy evaluation score;
generate, by the first artificial intelligence, a third response to the second response using the updated first response generation module based on the second empathy evaluation score, the second diversity-empathy evaluation score, and the second response;
and
provide, by the first artificial intelligence, the third response to the second artificial intelligence.

14 more elements

A method for generating an improved dialogue between artificial intelligences, the method comprising:

receiving, by one or more processors executing a first artificial intelligence, an initial input from a user;
generating, by the one or more processors executing the first artificial intelligence, a first response to the initial input using a first response generation module;
providing, by the one or more processors executing the first artificial intelligence, the first response to a second artificial intelligence;
determining, by the one or more processors executing the second artificial intelligence, a first empathy evaluation score for the first response using a first deep learning model trained with first data identifying empathy to be learned and empathy not to be learned;
determining, by the one or more processors executing the second artificial intelligence, a first diversity-empathy evaluation score for the first response using a second deep learning model trained with second data identifying whether to generate an emotional conversation or a diverse conversation;
updating, by the one or more processors executing the second artificial intelligence, a second response generation module based on the first empathy evaluation score and the first diversity-empathy evaluation score;
generating, by the one or more processors executing the second artificial intelligence, a second response to the first response using the updated second response generation module based on the first empathy evaluation score, the first diversity-empathy evaluation score, and the first response;
providing, by the one or more processors executing the second artificial intelligence, the second response to the first artificial intelligence;
determining, by the one or more processors executing the first artificial intelligence, a second empathy evaluation score for the second response using a third deep learning model trained with third data identifying empathy to be learned and empathy not to be learned;
determining, by the one or more processors executing the first artificial intelligence, a second diversity-empathy evaluation score for the second response using a fourth deep learning model trained with fourth data identifying whether to generate an emotional conversation or a diverse conversation;
updating, by the one or more processors executing the first artificial intelligence, the first response generation module based on the second empathy evaluation score and the second diversity-empathy evaluation score;
generating, by the one or more processors executing the first artificial intelligence, a third response to the second response using the updated first response generation module based on the second empathy evaluation score, the second diversity-empathy evaluation score, and the second response;
and
providing, by the one or more processors executing the first artificial intelligence, the third response to the second artificial intelligence.

11 more elements

Dependencies are inferred from the claim text.