Timezone: »

Averaged-DQN: Variance Reduction and Stabilization for Deep Reinforcement Learning
Oron Anschel · Nir Baram · Nahum Shimkin

Tue Aug 08 09:42 PM -- 10:00 PM (PDT) @ C4.5

Instability and variability of Deep Reinforcement Learning (DRL) algorithms tend to adversely affect their performance. Averaged-DQN is a simple extension to the DQN algorithm, based on averaging previously learned Q-values estimates, which leads to a more stable training procedure and improved performance by reducing approximation error variance in the target values. To understand the effect of the algorithm, we examine the source of value function estimation errors and provide an analytical comparison within a simplified model. We further present experiments on the Arcade Learning Environment benchmark that demonstrate significantly improved stability and performance due to the proposed extension.

Author Information

Oron Anschel (Technion)
Nir Baram (Technion - Israel Institute of Technology)
Nahum Shimkin (Technion)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors