Reinforcement Learning
First-Explore, then Exploit: Meta-Learning to Solve Hard Exploration-Exploitation Trade-Offs
·3099 words·15 mins·
loading
·
loading
Machine Learning
Reinforcement Learning
🏢 Department of Computer Science, University of British Columbia
Meta-RL agents often fail to explore effectively in environments where optimal behavior requires sacrificing immediate rewards for greater future gains. First-Explore, a novel method, tackles this by…
Finding good policies in average-reward Markov Decision Processes without prior knowledge
·1896 words·9 mins·
loading
·
loading
Machine Learning
Reinforcement Learning
🏢 Inria
First near-optimal reinforcement learning algorithm achieving best policy identification in average-reward MDPs without prior knowledge of complexity.
Federated Natural Policy Gradient and Actor Critic Methods for Multi-task Reinforcement Learning
·1640 words·8 mins·
loading
·
loading
Machine Learning
Reinforcement Learning
🏢 Carnegie Mellon University
This paper introduces federated natural policy gradient and actor-critic methods achieving near dimension-free global convergence for decentralized multi-task reinforcement learning, a significant bre…
Federated Ensemble-Directed Offline Reinforcement Learning
·2286 words·11 mins·
loading
·
loading
Machine Learning
Reinforcement Learning
🏢 Department of Electrical and Computer Engineering, Texas A&M University
FEDORA, a novel algorithm, enables high-quality policy learning in federated offline reinforcement learning by leveraging the collective wisdom of diverse client datasets without data sharing.
Fast TRAC: A Parameter-Free Optimizer for Lifelong Reinforcement Learning
·2979 words·14 mins·
loading
·
loading
Machine Learning
Reinforcement Learning
🏢 Harvard University
TRAC: a parameter-free optimizer conquering lifelong RL’s plasticity loss!
Extensive-Form Game Solving via Blackwell Approachability on Treeplexes
·2500 words·12 mins·
loading
·
loading
Reinforcement Learning
🏢 Columbia University
First algorithmic framework for Blackwell approachability on treeplexes, enabling stepsize-invariant EFG solvers with state-of-the-art convergence rates.
Exploring the Edges of Latent State Clusters for Goal-Conditioned Reinforcement Learning
·3530 words·17 mins·
loading
·
loading
AI Generated
Machine Learning
Reinforcement Learning
🏢 Rutgers University
CE2: A new goal-directed exploration algorithm for efficient reinforcement learning in unknown environments, prioritizing accessible frontier goals via latent state clustering.
Exploration by Learning Diverse Skills through Successor State Representations
·2767 words·13 mins·
loading
·
loading
AI Generated
Machine Learning
Reinforcement Learning
🏢 ISAE-Supaero
LEADS: a novel algorithm learning diverse skills through successor state representations for robust exploration in reward-free environments.
Exploiting the Replay Memory Before Exploring the Environment: Enhancing Reinforcement Learning Through Empirical MDP Iteration
·2926 words·14 mins·
loading
·
loading
Machine Learning
Reinforcement Learning
🏢 Department of Computing Science and Amii, University of Alberta
Boost RL performance by solving a series of simplified MDPs before tackling the complex real-world one!
Exclusively Penalized Q-learning for Offline Reinforcement Learning
·2010 words·10 mins·
loading
·
loading
Reinforcement Learning
🏢 UNIST
EPQ, a novel offline RL algorithm, significantly reduces underestimation bias by selectively penalizing states prone to errors, improving performance over existing methods.
Excluding the Irrelevant: Focusing Reinforcement Learning through Continuous Action Masking
·2172 words·11 mins·
loading
·
loading
Machine Learning
Reinforcement Learning
🏢 Technical University of Munich
Boost RL efficiency in continuous action spaces by masking irrelevant actions using three novel continuous action masking methods!
Episodic Future Thinking Mechanism for Multi-agent Reinforcement Learning
·2505 words·12 mins·
loading
·
loading
Machine Learning
Reinforcement Learning
🏢 Soongsil University
RL agents make better decisions by simulating future scenarios, considering diverse agent behaviors, and using character inference for improved decision-making.
Entropy-regularized Diffusion Policy with Q-Ensembles for Offline Reinforcement Learning
·3343 words·16 mins·
loading
·
loading
AI Generated
Machine Learning
Reinforcement Learning
🏢 Uppsala University
Entropy-regularized diffusion policy with Q-ensembles achieves state-of-the-art offline reinforcement learning by tackling overestimation of Q-values and boosting exploration.
Ensemble sampling for linear bandits: small ensembles suffice
·310 words·2 mins·
loading
·
loading
Machine Learning
Reinforcement Learning
🏢 University of Oxford
Small ensembles in stochastic linear bandits achieve near-optimal regret; a rigorous analysis shows that ensemble size need only scale logarithmically with horizon.
Enhancing Robustness in Deep Reinforcement Learning: A Lyapunov Exponent Approach
·2344 words·12 mins·
loading
·
loading
Machine Learning
Reinforcement Learning
🏢 University of Glasgow
Deep RL agents lack robustness; this paper enhances their resilience by implementing Maximal Lyapunov Exponent regularisation in the Dreamer V3 architecture, thus improving real-world applicability.
Enhancing Efficiency of Safe Reinforcement Learning via Sample Manipulation
·1946 words·10 mins·
loading
·
loading
Machine Learning
Reinforcement Learning
🏢 UC Berkeley
ESPO enhances safe RL efficiency by dynamically manipulating sample size based on reward-safety gradient conflicts, ensuring faster training and superior performance.
Enhancing Chess Reinforcement Learning with Graph Representation
·2930 words·14 mins·
loading
·
loading
AI Generated
Machine Learning
Reinforcement Learning
🏢 Kyoto University
AlphaGateau: a novel Graph Neural Network architecture outperforms previous chess AI models by leveraging graph representations for faster training and superior generalization to different board sizes…
Enabling Adaptive Agent Training in Open-Ended Simulators by Targeting Diversity
·2923 words·14 mins·
loading
·
loading
Machine Learning
Reinforcement Learning
🏢 University of Southern California
DIVA: Evolutionary task generation for robust, adaptable AI agents in complex simulators.
Efficient Reinforcement Learning by Discovering Neural Pathways
·3467 words·17 mins·
loading
·
loading
Machine Learning
Reinforcement Learning
🏢 McGill University
Discover efficient neural pathways for reinforcement learning; drastically reducing model size and energy consumption without sacrificing performance.
Efficient Recurrent Off-Policy RL Requires a Context-Encoder-Specific Learning Rate
·3209 words·16 mins·
loading
·
loading
Machine Learning
Reinforcement Learning
🏢 Nanjing University
Recurrent off-policy RL, while robust, suffers from training instability. RESEL, a novel algorithm, solves this by using a context-encoder-specific learning rate, significantly improving stability an…