GitPlanet

Cheap and reliable Node.js hosting starts at $3/month, and $1/month static HTML hosting

Created with love in Canada, visit hostnodejs.com today

Feel like to post an Ad? Learn Details
All Projects → Kaixhin → Noisynet A3c

Kaixhin / Noisynet A3c

Licence: mit
Noisy Networks for Exploration

Programming Languages

python
139335 projects - #7 most used programming language

Labels

deep-learningdeep-reinforcement-learning

Projects that are alternatives of or similar to Noisynet A3c

Adnet
Attention-guided CNN for image denoising(Neural Networks,2020)
Stars: ✭ 135 (-21.05%)
Mutual labels:  deep-reinforcement-learning
Scalphagozero
An independent implementation of DeepMind's AlphaGoZero in Scala, using Deeplearning4J (DL4J)
Stars: ✭ 144 (-15.79%)
Mutual labels:  deep-reinforcement-learning
Airsim
Open source simulator for autonomous vehicles built on Unreal Engine / Unity, from Microsoft AI & Research
Stars: ✭ 12,528 (+7226.32%)
Mutual labels:  deep-reinforcement-learning
Deep Qlearning Agent For Traffic Signal Control
A framework where a deep Q-Learning Reinforcement Learning agent tries to choose the correct traffic light phase at an intersection to maximize traffic efficiency.
Stars: ✭ 136 (-20.47%)
Mutual labels:  deep-reinforcement-learning
Deep Learning Papers Reading Roadmap
深度学习论文阅读路线图
Stars: ✭ 142 (-16.96%)
Mutual labels:  deep-reinforcement-learning
Minimalrl
Implementations of basic RL algorithms with minimal lines of codes! (pytorch based)
Stars: ✭ 2,051 (+1099.42%)
Mutual labels:  deep-reinforcement-learning
Keras Rl2
Reinforcement learning with tensorflow 2 keras
Stars: ✭ 134 (-21.64%)
Mutual labels:  deep-reinforcement-learning
Deep Reinforcement Learning Algorithms
31 projects in the framework of Deep Reinforcement Learning algorithms: Q-learning, DQN, PPO, DDPG, TD3, SAC, A2C and others. Each project is provided with a detailed training log.
Stars: ✭ 167 (-2.34%)
Mutual labels:  deep-reinforcement-learning
Pcgrad
Code for "Gradient Surgery for Multi-Task Learning"
Stars: ✭ 144 (-15.79%)
Mutual labels:  deep-reinforcement-learning
Spinning Up Basic
Basic versions of agents from Spinning Up in Deep RL written in PyTorch
Stars: ✭ 155 (-9.36%)
Mutual labels:  deep-reinforcement-learning
Machine Learning And Data Science
This is a repository which contains all my work related Machine Learning, AI and Data Science. This includes my graduate projects, machine learning competition codes, algorithm implementations and reading material.
Stars: ✭ 137 (-19.88%)
Mutual labels:  deep-reinforcement-learning
D3rlpy
An offline deep reinforcement learning library
Stars: ✭ 139 (-18.71%)
Mutual labels:  deep-reinforcement-learning
Go Bot Drl
Goal-Oriented Chatbot trained with Deep Reinforcement Learning
Stars: ✭ 149 (-12.87%)
Mutual labels:  deep-reinforcement-learning
Policy Gradient
Minimal Monte Carlo Policy Gradient (REINFORCE) Algorithm Implementation in Keras
Stars: ✭ 135 (-21.05%)
Mutual labels:  deep-reinforcement-learning
Gym Carla
An OpenAI gym wrapper for CARLA simulator
Stars: ✭ 164 (-4.09%)
Mutual labels:  deep-reinforcement-learning
Ml Agents
Unity Machine Learning Agents Toolkit
Stars: ✭ 12,134 (+6995.91%)
Mutual labels:  deep-reinforcement-learning
Baby A3c
A high-performance Atari A3C agent in 180 lines of PyTorch
Stars: ✭ 144 (-15.79%)
Mutual labels:  deep-reinforcement-learning
2048 Deep Reinforcement Learning
Trained A Convolutional Neural Network To Play 2048 using Deep-Reinforcement Learning
Stars: ✭ 169 (-1.17%)
Mutual labels:  deep-reinforcement-learning
Accel Brain Code
The purpose of this repository is to make prototypes as case study in the context of proof of concept(PoC) and research and development(R&D) that I have written in my website. The main research topics are Auto-Encoders in relation to the representation learning, the statistical machine learning for energy-based models, adversarial generation networks(GANs), Deep Reinforcement Learning such as Deep Q-Networks, semi-supervised learning, and neural network language model for natural language processing.
Stars: ✭ 166 (-2.92%)
Mutual labels:  deep-reinforcement-learning
Awesome Deep Neuroevolution
A collection of Deep Neuroevolution resources or evolutionary algorithms applying in Deep Learning (constantly updating)
Stars: ✭ 150 (-12.28%)
Mutual labels:  deep-reinforcement-learning
View All Similar Projects ➔

NoisyNet-A3C

MIT License

NoisyNet [1] (LSTM) asynchronous advantage actor-critic (A3C) [2] on the CartPole-v1 environment. This repo has a minimalistic design and a classic control environment to enable quick investigation of different hyperparameters.

Run with python main.py <options>. Entropy regularisation can still be added by setting --entropy-weight <value>, but it is 0 by default. Run with --no-noise to run normal A3C (without noisy linear layers).

Requirements

To install all dependencies with Anaconda run conda env create -f environment.yml and use source activate noisynet to activate the environment.

Results

NoisyNet-A3C

On the whole, NoisyNet-A3C tends to be better than A3C (with or without entropy regularisation). There seems to be more variance, with both good and poor runs, probably due to "deep" exploration.

Good-NoisyNet-A3C

Bad-NoisyNet-A3C

NoisyNet-A3C is perhaps even more prone to performance collapses than normal A3C. Many deep reinforcement learning algorithms are still prone to this.

Collapse-NoisyNet-A3C

A3C (no entropy regularisation)

A3C without entropy regularisation usually performs poorly.

A3C

A3C (entropy regularisation with β = 0.01)

A3C with entropy regularisation usually performs a bit better than A3C without entropy regularisation, and also poor runs of NoisyNet-A3C. The performance tends to be significantly worse than the best NoisyNet-A3C runs.

A3C-entropy

Note that due to the nondeterminism introduced by asynchronous agents, different runs on even the same seed can produce different results, and hence the results presented are only single samples of the performance of these algorithms. Interestingly, the general observations above seem to hold even when increasing the number of processes (experiments were repeated with 16 processes). These algorithms are still sensitive to the choice of hyperparameters, and will need to be tuned extensively to get good performance on other domains.

Acknowledgements

References

[1] Noisy Networks for Exploration
[2] Asynchronous Methods for Deep Reinforcement Learning

Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].