Dueling Network Architectures for Deep Reinforcement Learning

, , , , ,  -

Publications: arXiv Add/Edit

Abstract: Add/Edit

In recent years there have been many successes of using deep representations in reinforcement learning. Still, many of these applications use conventional architectures, such as convolutional networks, LSTMs, or auto-encoders. In this paper, we present a new neural network architecture for model-free reinforcement learning. Our dueling network represents two separate estimators: one for the state value function and one for the state-dependent action advantage function. The main benefit of this factoring is to generalize learning across actions without imposing any change to the underlying reinforcement learning algorithm. Our results show that this architecture leads to better policy evaluation in the presence of many similar-valued actions. Moreover, the dueling architecture enables our RL agent to outperform the state-of-the-art on the Atari 2600 domain.

Keywords: Add/Edit


Code Links

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Deep Learning papers reading roadmap for anyone who are eager to learn this amazing tech!

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Line plot extraction using Reinforcement learning (Dueling DQN) using Keras

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Tensorflow Implementation for "Noisy network for exploration"

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

A deep reinforcement learning library in tensorflow

0

Languages: Jupyter Notebook Add/Edit

Libraries: Add/Edit

Description: Add/Edit

CS 294: Deep Reinforcement Learning, Spring 2017 Berkeley

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Atari - Deep Reinforcement Learning algorithms in TensorFlow

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

AI learning from visual input using ViZDoom environment.

0

Languages: AppleScript Add/Edit

Libraries: Add/Edit

Description: Add/Edit

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

TensorFlow implementation of Deep RL (Reinforcement Learning) papers based on deep Q-learning (DQN)

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Trading Gym is an open source project for the development of reinforcement learning algorithms in the context of trading.

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Tensorflow + OpenAI Gym implementation of Deep Q-Network (DQN), Double DQN (DDQN), Dueling Network and Deep Deterministic Policy Gradient (DDPG)

0

Languages: ASP Add/Edit

Libraries: Add/Edit

Description: Add/Edit

An RL agent that implements Dueling DDQN algorithm with Prioritized Experience Replay

0

Languages: Jupyter Notebook Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Implementation of Reinforcement Learning Algorithms. Python, OpenAI Gym, Tensorflow. Exercises and Solutions to accompany Sutton's Book and David Silver's course.

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Chainer implementation of Double DQN + Dueling Network

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Deep Reinforcement Learning (DQN) based Self Driving Car Control with Vehicle Simulator

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Implementation of a Dueling Double Deep Q-value Network with tensorflow

0
Github: Kyushik/DRL

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Repository for codes of 'Deep Reinforcement Learning'

0

Languages: Jupyter Notebook Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Decoders for fault tolerant quantum computation via deepQ reinforcement learning

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Deep reinforcement learning in ViZDoom (using Tensorflow)

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Deep Reinforcement Learning framework based on TensorFlow and OpenAI Gym

0

Languages: ASP Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Reinforcement Learning agent using DDPG implementation for continous control of the Reacher environment.

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Deep Reinforcement Learning for Keras.

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Rainbow: Combining Improvements in Deep Reinforcement Learning

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Reinforcement Learning

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Duel_DDQN (Dueling Network Architectures + Double DQN) using Keras

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Self-Driving Truck in Euro Truck Simulator 2, trained via Reinforcement Learning

0

Languages: Lua Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Persistent advantage learning dueling double DQN for the Arcade Learning Environment

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

TensorFlow implementation of Deep Reinforcement Learning papers

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

[Review & Implementation]

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

rlcourse-march-31-nke001 created by GitHub Classroom

0
github.com: /FaboNo/DRLND

Languages: Jupyter Notebook Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Repository related to projects for the Udacity Deep Reinforcement Learning Nano Degree

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Highly modularized implementation of popular deep RL algorithms by PyTorch

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Deep Reinforcement Learning related works.

0

Languages: Roff Add/Edit

Libraries: Add/Edit

Description: Add/Edit

This project contains our source code of reinforcement learning compettion on halite.io

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Implementation of DQN for Super Mario Bros. 1-1

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

An experiment on using deep learning to play games

0

Languages: Python Add/Edit

Libraries: Add/Edit

Description: Add/Edit

Keras implementation of DQN on ViZDoom environment

0

Languages: ASP Add/Edit

Libraries: Add/Edit

Description: Add/Edit

DuelDQN with Pytorch

0