Deep Active Inference for Pixel-Based Discrete Control: Evaluation on the Car Racing Problem
Abstract
Despite the potential of active inference for visual-based control, learning the model and the preferences (priors) while interacting with the environment is challenging. Here, we study the performance of a deep active inference (dAIF) agent on OpenAI's car racing benchmark, where there is no access to the car's state. The agent learns to encode the world's state from high-dimensional input through unsupervised representation learning. State inference and control are learned end-to-end by optimizing the expected free energy. Results show that our model achieves comparable performance to deep Q-learning. However, vanilla dAIF does not reach state-of-the-art performance compared to other world model approaches. Hence, we discuss the current model implementation's limitations and potential architectures to overcome them.
Cite
@article{arxiv.2109.04155,
title = {Deep Active Inference for Pixel-Based Discrete Control: Evaluation on the Car Racing Problem},
author = {Niels van Hoeffelen and Pablo Lanillos},
journal= {arXiv preprint arXiv:2109.04155},
year = {2021}
}
Comments
2nd International Workshop on Active Inference IWAI2021, European Conference on Machine Learning (ECML/PCKDD 2021)