Timezone: »

Revisiting Fundamentals of Experience Replay
William Fedus · Prajit Ramachandran · Rishabh Agarwal · Yoshua Bengio · Hugo Larochelle · Mark Rowland · Will Dabney

Wed Jul 15 11:00 AM -- 11:45 AM & Wed Jul 15 10:00 PM -- 10:45 PM (PDT) @ Virtual #None

Experience replay is central to off-policy algorithms in deep reinforcement learning (RL), but there remain significant gaps in our understanding. We therefore present a systematic and extensive analysis of experience replay in Q-learning methods, focusing on two fundamental properties: the replay capacity and the ratio of learning updates to experience collected (replay ratio). Our additive and ablative studies upend conventional wisdom around experience replay — greater capacity is found to substantially increase the performance of certain algorithms, while leaving others unaffected. Counterintuitively we show that theoretically ungrounded, uncorrected n-step returns are uniquely beneficial while other techniques confer limited benefit for sifting through larger memory. Separately, by directly controlling the replay ratio we contextualize previous observations in the literature and empirically measure its importance across a variety of deep RL algorithms. Finally, we conclude by testing a set of hypotheses on the nature of these performance benefits.

Author Information

William Fedus (Google Brain/Mila)
Prajit Ramachandran (Google)
Rishabh Agarwal (Google Research, Brain Team)
Yoshua Bengio (Montreal Institute for Learning Algorithms)
Hugo Larochelle (Google Brain)
Mark Rowland (DeepMind)
Will Dabney (DeepMind)

More from the Same Authors