paper-with-me

Papers

On-Policy Trust Region Policy Optimisation with Replay Buffers

2019-01-18 · ICLR 2019 5 · Dmitry Kangin, Nicolas Pugeault

Building upon the recent success of deep reinforcement learning methods, we investigate the possibility of on-policy reinforcement learning improvement by reusing the data from several consecutive policies. On-policy methods bring many benefits, such as ability to evaluate each resulting policy. However, they usually discard all the information about the policies which existed before. In this work, we propose adaptation of the replay buffer concept, borrowed from the off-policy learning setting, to create the method, combining advantages of on- and off-policy learning. To achieve this, the proposed algorithm generalises the $Q$-, value and advantage functions for data from multiple policies. The method uses trust region optimisation, while avoiding some of the common problems of the algorithms such as TRPO or ACKTR: it uses hyperparameters to replace the trust region selection heuristics, as well as the trainable covariance matrix instead of the fixed one. In many cases, the method not only improves the results comparing to the state-of-the-art trust region on-policy learning algorithms such as PPO, ACKTR and TRPO, but also with respect to their off-policy counterpart DDPG.

📄 PDF Abstract BibTeX arXiv:1901.06212

Code (2)

dkangin/baselines 공식 구현 tf
dkangin/baselines/tree/master/baselines/trpo_replay tf

Tasks

Continuous ControlDeep Reinforcement LearningPolicy Gradient Methodsreinforcement-learningReinforcement LearningReinforcement Learning (RL)

Methods 이 논문이 사용한 방법론

Entropy Regularization 설명 없음
Weight Decay 설명 없음
Convolution A convolution is a type of matrix operation, consisting of a kernel, a small matrix of weights, that slides over input data performing element-wise multiplication with the…
PPO Proximal Policy Optimization, or PPO, is a policy gradient method for reinforcement learning. The motivation was to have an algorithm with the data efficiency and reliable…
Adam 설명 없음
Dense Connections Dense Connections, or Fully Connected Connections, are a type of layer in a deep neural network that use a linear operation where every input is connected to every output…
Batch Normalization 설명 없음
ReLU How Do I Communicate to Expedia? How Do I Communicate to Expedia? – Call ☎️ +1-(888) 829 (0881) or +1-805-330-4056 or +1-805-330-4056 for Live Support & Special Travel…

Similar Papers 제목 키워드 기반

Sample-efficient Actor-Critic Reinforcement Learning with Supervised Data for Dialogue Management

2017-07-01 · WS 2017 8 · Pei-Hao Su, Pawel Budzianowski, Stefan Ultes, Milica Gasic 외

Deep reinforcement learning (RL) methods have significant potential for dialogue policy optimisation. However, they suffer from a poor performance in the early stages of learning. This is especially problematic for on-li…

Deep Reinforcement LearningDialogue ManagementManagementreinforcement-learning+2

Trust Region Policy Optimisation in Multi-Agent Reinforcement Learning

2021-09-23 · ICLR 2022 4 · Jakub Grudzien Kuba, Ruiqing Chen, Muning Wen, Ying Wen 외

Trust region methods rigorously enabled reinforcement learning (RL) agents to learn monotonically improving policies, leading to superior performance on a variety of tasks. Unfortunately, when it comes to multi-agent rei…

LEMMAMuJoCoMulti-agent Reinforcement Learningreinforcement-learning+2

Remember and Forget for Experience Replay

2018-07-16 · ICLR 2019 5 · Guido Novati, Petros Koumoutsakos

Experience replay (ER) is a fundamental component of off-policy deep reinforcement learning (RL). ER recalls experiences from past iterations to compute gradient estimates for the current policy, increasing data-efficien…

Deep Reinforcement LearningPolicy Gradient MethodsQ-LearningReinforcement Learning+1

Efficient Off-Policy Safe Reinforcement Learning Using Trust Region Conditional Value at Risk

2023-12-01 · Dohyeong Kim, Songhwai Oh

This paper aims to solve a safe reinforcement learning (RL) problem with risk measure-based constraints. As risk measures, such as conditional value at risk (CVaR), focus on the tail distribution of cost signals, constra…

Reinforcement Learning (RL)Safe Reinforcement Learning

Maximum Entropy On-Policy Actor-Critic via Entropy Advantage Estimation

2024-07-25 · Jean Seong Bjorn Choe, Jong-Kook Kim

Entropy Regularisation is a widely adopted technique that enhances policy optimisation performance and stability. A notable form of entropy regularisation is augmenting the objective with an entropy term, thereby simulta…

MuJoCo