Back to results
US 2020/0026959 A1Published application
Electronic apparatus and learning method of electronic apparatus
AssigneeSamsung Electronics Co., Ltd.
InventorsDonghyeon LEE, Heriberto Cuayahuitl, Seonghan Ryu, Sungja Choi
Life of this patent
Estimating the term…
Abstract
Artificial intelligence for machine learning to provide an optimized response sentence in reply to an input sentence.
Claims18
A method of an electronic apparatus providing a response sentence in reply to an input sentence of a user, the method comprising:
- obtaining a plurality of conversation sentences;
- obtaining a plurality of clusters for the plurality of conversation sentences based on similarity between the plurality of conversation sentences;
- based on the input sentence, providing the input sentence as input to an artificial intelligence model trained to provide the response sentence in reply to the input sentence and providing a cluster from among the plurality of clusters as output from the artificial intelligence model;
- providing a conversation sentence from among the plurality of conversation sentences included in the cluster as the response sentence in reply to the input sentence of the user;
- obtaining a reward value for the response sentence;
- and
- updating weight values of the artificial intelligence model based on the reward value to optimize the response sentence provided in reply to the input sentence.
4 more elements
An electronic apparatus comprising:
- a memory configured to store computer-readable instructions;
- and
- a processor configured to execute the computer-readable instructions control the electronic apparatus to:
- obtain a plurality of conversation sentences;
- obtain a plurality of clusters for the plurality of conversation sentences based on similarity between the plurality of conversation sentences;
- based on an input sentence being input from a user, provide the input sentence as input to an artificial intelligence model trained to provide a response sentence in reply to the input sentence and provide a cluster from among the plurality of clusters as output from the artificial intelligence model;
- provide a conversation sentence from among the plurality of conversation sentences included in the cluster as the response sentence in reply to the input sentence of the user;
- obtain a reward value for the response sentence;
- and
- update weight values of the artificial intelligence model based on the reward value to optimize the response sentence provided in reply to the input sentence.
7 more elements
Dependencies are inferred from the claim text.