GithubHelp home page GithubHelp logo

cun-bjy / rl-paper-review Goto Github PK

View Code? Open in Web Editor NEW
38.0 5.0 6.0 6.56 MB

road-map & paper review for Reinforcement Learning

License: MIT License

policy-gradient rl reinforcement-learning paper-review ddpg ppo gae trpo td3 sac

rl-paper-review's Introduction

rl-paper-review's People

Contributors

cun-bjy avatar

Stargazers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

Watchers

 avatar  avatar  avatar  avatar  avatar

rl-paper-review's Issues

[TD3] 4.1 Overestimation Bias in Actor-Critic

1. 충분히 작은 error <img src="https://latex.codecogs.com/gif.latex?\epsilon_1" title="\epsilon_1" />이 존재(<img src="https://latex.codecogs.com/gif.latex?\alpha\ge\epsilon_1" title="\alpha\ge\epsilon_1" />)한다고 할 때, <img src="https://latex.codecogs.com/gif.latex?\pi_\text{approx}" title="\pi_\text{approx}" />의 approximate value는 <img src="https://latex.codecogs.com/gif.latex?\pi_\text{true}" title="\pi_\text{true}" /> approximate value보다 작거나 같다.
<img src="../img/td3_14.png"/>
2. 마찬가지로 충분히 작은 error <img src="https://latex.codecogs.com/gif.latex?\epsilon_2" title="\epsilon_2" />이 존재(<img src="https://latex.codecogs.com/gif.latex?\alpha\ge\epsilon_2" title="\alpha\ge\epsilon_2" />)한다고 할 때, <img src="https://latex.codecogs.com/gif.latex?\pi_\text{true}" title="\pi_\text{true}" />의 true value는 <img src="https://latex.codecogs.com/gif.latex?\pi_\text{approx}" title="\pi_\text{approx}" />의 true value보다 작거나 같다.
<img src="../img/td3_15.png"/>

  • 97줄: 논문에 "the approximate value of π_approx will be bounded below by the approximate value of π_true"라고 되어있으므로 해석을 "크거나 같다"라고 해야 할 것 같습니다. 그래야 식의 의미와도 일치한다고 생각합니다.
  • 101줄: 마찬가지입니다.

LaTex 수식 인식문제

Typora를 통해서는 인식이 잘 되지만 GitHub md 에서는 LaTex 수식이 인식되지않아 convert 모듈을 이용해아함.

현재 컴파일 에러중..

[TD3] 5. Addressing Variance

기존 Sutton & Barto의 논문에서 밝혀졌듯이 learning speed를 줄이는 것은 variance를 줄일수는 있지만 실제로 performance에 영향을 줄 수 있다. 해당 논문에서는 **estimation error 최소화의 중요성**을 다시금 강조하고, **target network와 estimation error를 연관지어 variance를 줄일 수 있는 개선된 actor-critic 학습 과정을 제시**하려한다.

  • Sutton & Barto, 1998
    • 논문이 아니고 책입니다^^ 강화학습의 교과서라고 불리지요ㅎㅎ 2018년엔 2쇄본이 나왔습니다 (https://a.co/d/fv54CBn)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.