Advantage estimation method and device, electronic equipment and storage medium

Abstract

The invention provides an advantage estimation method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring a current environment state; inputting the current environment state into the advantage estimation model to obtain an advantage action obtained by carrying out advantage estimation on the basis of the current environment state by the advantage estimation model, wherein the advantage estimation model is obtained by training based on a teaching data set and a behavior cloning model, the teaching data set comprises a sample environment state and a sample action corresponding to the sample environment state, and the behavior clone model is obtained based on training of the teaching data set. According to the invention, the advantage estimation model is trained based on the teaching data set and the behavior clone model, through the self-adaptive behavior clone model, the teaching data are fully utilized, expert experience in the historical teaching data is automatically mined, adverse effects possibly brought by incomplete teaching data are avoided, the advantage estimation performance of the advantage estimation model is enhanced, and the advantage estimation accuracy in a complex scene is improved.

Description

technical field [0001] The present invention relates to the technical field of reinforcement learning, in particular to an advantage estimation method, device, electronic equipment and storage medium. Background technique [0002] Deep Reinforcement Learning (DRL) has made great progress in recent years and is widely used in decision-making scenarios such as video games and board games. With the help of deep learning's powerful feature extraction and function fitting capabilities, reinforcement learning subjects can directly extract and learn feature knowledge from original input data (such as game images), and then use traditional reinforcement learning algorithms to learn decision-making control based on the extracted feature information. Strategies without manually extracting or learning features based on rules and heuristics. [0003] But currently, deep reinforcement learning techniques are still not practical for applications in solving complex decision-making control...

AI Technical Summary

Problems solved by technology

[0006] The present invention provides an advantage estimation method, device, electronic equipment and

Images