Analytically Tractable Bayesian Deep Q-Learning
This work addresses the problem of scaling Bayesian methods in reinforcement learning for researchers and practitioners, offering a more efficient and hyperparameter-light alternative to gradient-based optimization, though it is incremental as it builds on existing Q-learning and TAGI frameworks.
The paper tackles the challenge of scaling Bayesian deep learning methods to complex reinforcement learning benchmarks like Atari games by adapting temporal difference Q-learning to be compatible with tractable approximate Gaussian inference (TAGI), enabling closed-form analytical inference for neural network parameters. The result shows that TAGI achieves performance comparable to backpropagation-trained networks while using fewer hyperparameters and no gradient-based optimization.
Reinforcement learning (RL) has gained increasing interest since the demonstration it was able to reach human performance on video game benchmarks using deep Q-learning (DQN). The current consensus for training neural networks on such complex environments is to rely on gradient-based optimization. Although alternative Bayesian deep learning methods exist, most of them still rely on gradient-based optimization, and they typically do not scale on benchmarks such as the Atari game environment. Moreover none of these approaches allow performing the analytical inference for the weights and biases defining the neural network. In this paper, we present how we can adapt the temporal difference Q-learning framework to make it compatible with the tractable approximate Gaussian inference (TAGI), which allows learning the parameters of a neural network using a closed-form analytical method. Throughout the experiments with on- and off-policy reinforcement learning approaches, we demonstrate that TAGI can reach a performance comparable to backpropagation-trained networks while using fewer hyperparameters, and without relying on gradient-based optimization.