Reinforcement Learning
Trust Region-Guided Proximal Policy Optimization
Yuhui Wang, Hao He, Xiaoyang Tan, Yaozhong Gan
However, the first-order optimizer is not very accurate for curved areas.
Dual Generator Offline Reinforcement Learning
Such promise is especially appealing in domains where data collection is expensive or dangerous, but large amounts of data may already exists (e.g., robotics, autonomous driving, task-oriented dialog systems).
Gossip-based Actor-Learner Architectures for Deep Reinforcement Learning
Mahmoud ("Mido") Assran, Joshua Romoff, Nicolas Ballas, Joelle Pineau, Mike Rabbat
Neural Information Processing Systems http://nips.cc/
Explicable Policy Search
Human teammates often form conscious and subconscious expectations of each other during interaction. Teaming success is contingent on whether such expectations can be met. Similarly, for an intelligent agent to operate beside a human, it must consider the human's expectation of its behavior. Disregarding such expectations can lead to the loss of trust and degraded team performance. A key challenge here is that the human's expectation may not align with the agent's
Regret Minimization for Reinforcement Learning by Evaluating the Optimal Bias Function
Therefore, there is a trade-off between exploration and exploitation, i.e., taking actions we have not learned accurately enough and taking actions which
Operator Splitting Value Iteration
OS-VI achieves a much faster convergence rate when the model is accurate enough.
Double Check Y our State Before Trusting It: Confidence-Aware Bidirectional Offline Model-Based Imagination
However, the imagined transitions may be inaccurate, thus downgrading the performance of the underlying offline RL method.
Exploit Reward Shifting in Value-Based Deep-RL: Optimistic Curiosity-Based Exploration and Conservative Exploitation via Linear Reward Shaping Hao Sun 1 Lei Han 2 Rui Y ang
In this work, we study the simple yet universally applicable case of reward shaping in value-based Deep Reinforcement Learning (DRL).