Symmetric Q-learning: Reducing Skewness of Bellman Error in Online Reinforcement Learning
This work addresses a specific bottleneck in reinforcement learning for continuous control tasks, offering an incremental improvement to existing methods.
The paper tackled the problem of skewed error distributions in value function estimation in deep reinforcement learning, which violates the Gaussian assumption of least squares methods, by proposing Symmetric Q-learning that adds synthetic noise to target values to achieve a Gaussian error distribution, resulting in improved sample efficiency on MuJoCo benchmark tasks.
In deep reinforcement learning, estimating the value function to evaluate the quality of states and actions is essential. The value function is often trained using the least squares method, which implicitly assumes a Gaussian error distribution. However, a recent study suggested that the error distribution for training the value function is often skewed because of the properties of the Bellman operator, and violates the implicit assumption of normal error distribution in the least squares method. To address this, we proposed a method called Symmetric Q-learning, in which the synthetic noise generated from a zero-mean distribution is added to the target values to generate a Gaussian error distribution. We evaluated the proposed method on continuous control benchmark tasks in MuJoCo. It improved the sample efficiency of a state-of-the-art reinforcement learning method by reducing the skewness of the error distribution.