paper-with-me

Papers

Evolution Strategies as a Scalable Alternative to Reinforcement Learning

2017-03-10 · Tim Salimans, Jonathan Ho, Xi Chen, Szymon Sidor, Ilya Sutskever

We explore the use of Evolution Strategies (ES), a class of black box optimization algorithms, as an alternative to popular MDP-based RL techniques such as Q-learning and Policy Gradients. Experiments on MuJoCo and Atari show that ES is a viable solution strategy that scales extremely well with the number of CPUs available: By using a novel communication strategy based on common random numbers, our ES implementation only needs to communicate scalars, making it possible to scale to over a thousand parallel workers. This allows us to solve 3D humanoid walking in 10 minutes and obtain competitive results on most Atari games after one hour of training. In addition, we highlight several advantages of ES as a black box optimization technique: it is invariant to action frequency and delayed rewards, tolerant of extremely long horizons, and does not need temporal discounting or value function approximation.

📄 PDF Abstract BibTeX arXiv:1703.03864

Code (23)

openai/evolution-strategies-starter 공식 구현 tf
AnshMittal1811/AugmentedRandomSearch
FlixCoder/rust-es-optimizer
MohammadAsadolahi/Gradient-Free-Reinforcement-Learning-solving-Openai-gym-LunarLanderV2-by-Evolution-Strategy-python/blob/main/README.md pytorch
ShangtongZhang/DistributedES pytorch
alisidd/Evolution-Strategies tf
aspk/space_battle
atgambardella/pytorch-es pytorch
cesch97/NeuroEvolution
czen88/qtrader tf
enajx/ES pytorch
evaboost/evaboost
fiberleif/evolution-strategies tf
neilsgp/RL-Algorithms
neuro-soup/es-torch jax
nnaisense/pgpelib pytorch
patniemeyer/ga-autoencoder pytorch
rasmusbergpalm/evostrat pytorch
ray-project/ray/tree/master/rllib
shakti365/nes
silvialuu/DRL-2018 pytorch
stavrosgreece/MultiAgentLearning pytorch
susanwe/ope_worldmodels

Tasks

Atari GamesMuJoCoQ-Learningreinforcement-learningReinforcement LearningReinforcement Learning (RL)

Methods 이 논문이 사용한 방법론

Q-Learning Q-Learning is an off-policy temporal difference control algorithm: $$Q\left(S\_{t}, A\_{t}\right) \leftarrow Q\left(S\_{t}, A\_{t}\right) + \alpha\left[R_{t+1} +…

Similar Papers 제목 키워드 기반

Importance Weighted Evolution Strategies

2018-11-12 · Víctor Campos, Xavier Giro-i-Nieto, Jordi Torres

Evolution Strategies (ES) emerged as a scalable alternative to popular Reinforcement Learning (RL) techniques, providing an almost perfect speedup when distributed across hundreds of CPU cores thanks to a reduced communi…

CPUreinforcement-learningReinforcement LearningReinforcement Learning (RL)

Instance Weighted Incremental Evolution Strategies for Reinforcement Learning in Dynamic Environments

2020-10-09 · Zhi Wang, Chunlin Chen, Daoyi Dong

Evolution strategies (ES), as a family of black-box optimization algorithms, recently emerge as a scalable alternative to reinforcement learning (RL) approaches such as Q-learning or policy gradient, and are much faster …

Incremental LearningQ-Learningreinforcement-learningReinforcement Learning (RL)+1

Guiding Evolutionary Strategies by Differentiable Robot Simulators

2021-10-01 · Vladislav Kurenkov, Bulat Maksudov

In recent years, Evolutionary Strategies were actively explored in robotic tasks for policy search as they provide a simpler alternative to reinforcement learning algorithms. However, this class of algorithms is often cl…

reinforcement-learningReinforcement Learning (RL)

Hard-Thresholding Meets Evolution Strategies in Reinforcement Learning

2024-05-02 · Chengqian Gao, William de Vazelhes, Hualin Zhang, Bin Gu 외

Evolution Strategies (ES) have emerged as a competitive alternative for model-free reinforcement learning, showcasing exemplary performance in tasks like Mujoco and Atari. Notably, they shine in scenarios with imperfect …

Decision MakingMuJoCoreinforcement-learningReinforcement Learning

Variance-Reduced Gradient Estimation via Noise-Reuse in Online Evolution Strategies

2023-04-21 · NeurIPS 2023 11 · Oscar Li, James Harrison, Jascha Sohl-Dickstein, Virginia Smith 외

Unrolled computation graphs are prevalent throughout machine learning but present challenges to automatic differentiation (AD) gradient estimation methods when their loss functions exhibit extreme local sensitivtiy, disc…