Dialogue information acquisition method and device, equipment and storage medium

Abstract

The invention discloses a dialogue information acquisition method and device, equipment and a storage medium, and relates to the technical field of artificial intelligence. The method comprises the following steps: acquiring first dialogue information; based on the first dialogue information, performing data processing through a first dialogue model to obtain second dialogue information; based on the first dialogue information and the second dialogue information, performing data processing through a first discriminator to obtain a first confidence score; and performing parameter updating on the first dialogue model based on the first confidence score. According to the method, the trained discriminator can detect the matching degree between the first dialogue information input into the first dialogue model and the second dialogue information output by the first dialogue model, the matching score serves as the reward value, each round of dialogue can obtain the reward value corresponding to the round of dialogue, the reward value generation density is improved, and therefore, the training efficiency of the first dialogue model obtained through reinforcement learning training is improved.

Description

technical field [0001] The present application relates to the technical field of artificial intelligence, and in particular to a dialog information acquisition method, device, equipment and storage medium. Background technique [0002] Intelligent dialogue is a series of related technologies that receive text information input by users and process it through artificial intelligence to respond to the text information input by users. [0003] In the related art, an intelligent dialogue system includes a task-oriented dialogue system and a non-task-oriented dialogue system (such as a chat robot). Task-oriented dialogue systems are usually trained through reinforcement learning. In the process of training task-oriented dialogue systems through reinforcement learning, when each dialogue is completed, it can be based on the dialogue of this dialogue. The number of rounds and the success of the dialogue, determine the reward function, and update the parameters of the model based o...

AI Technical Summary

Problems solved by technology

[0004] In the above technical solution, when the task-oriented dialogue system is trained by means of reinforcement learning, the dialogue reward can only be obtained according to the reward function after each dialogue is completed, resulting in relatively sparse training rewards and low training efficiency

Images