36 citations · 57 across the 12 of their papers we have counts for
10 papers · 1 filter
Trajectory-Aware Eligibility Traces for Off-Policy Reinforcement Learning
Brett Daley, Martha White, Christopher Amato +1
Off-policy learning from multistep returns is crucial for sample-efficient reinforcement learning, but counteracting off-policy bias without exacerbating variance is challenging. C…
Adaptive Tree Backup Algorithms for Temporal-Difference Reinforcement Learning
Brett Daley, Isaac Chan
Q() is a recently proposed temporal-difference learning method that interpolates between learning from expected backups and sampled backups. It has been shown that intermediate…
Improving the Efficiency of Off-Policy Reinforcement Learning by Accounting for Past Decisions
Brett Daley, Christopher Amato
Off-policy learning from multistep returns is crucial for sample-efficient reinforcement learning, particularly in the experience replay setting now commonly used with deep neural…
Virtual Replay Cache
Brett Daley, Christopher Amato
Return caching is a recent strategy that enables efficient minibatch training with multistep estimators (e.g. the λ-return) for deep reinforcement learning. By precomputing return…
Human-Level Control without Server-Grade Hardware
Brett Daley, Christopher Amato
Deep Q-Network (DQN) marked a major milestone for reinforcement learning, demonstrating for the first time that human-level control policies could be learned directly from raw visu…
Investigating Alternatives to the Root Mean Square for Adaptive Gradient Methods
Brett Daley, Christopher Amato
Adam is an adaptive gradient method that has experienced widespread adoption due to its fast and reliable training performance. Recent approaches have not offered significant impro…
Stratified Experience Replay: Correcting Multiplicity Bias in Off-Policy Reinforcement Learning
Brett Daley, Cameron Hickert, Christopher Amato
Deep Reinforcement Learning (RL) methods rely on experience replay to approximate the minibatched supervised learning setting; however, unlike supervised learning where access to l…
Contrasting Centralized and Decentralized Critics in Multi-Agent Reinforcement Learning
Xueguang Lyu, Yuchen Xiao, Brett Daley +1
Centralized Training for Decentralized Execution, where agents are trained offline using centralized information but execute in a decentralized manner online, has gained popularity…
Expectigrad: Fast Stochastic Optimization with Robust Convergence Properties
Brett Daley, Christopher Amato
Many popular adaptive gradient methods such as Adam and RMSProp rely on an exponential moving average (EMA) to normalize their stepsizes. While the EMA makes these methods highly r…
Reconciling -Returns with Experience Replay
Brett Daley, Christopher Amato
Modern deep reinforcement learning methods have departed from the incremental learning required for eligibility traces, rendering the implementation of the -return difficult in…