Distributional Active Inference
Abdullah Akgül, Gulcin Baykal, Manuel Haussmann, Mustafa Mert Çelikok, Melih Kandemir
Abstract
Optimal control of complex environments with robotic systems faces two complementary and intertwined challenges: efficient organization of sensory state information and far-sighted action planning. Because the reinforcement learning framework addresses only the latter, it tends to deliver sample-inefficient solutions. Active inference is the state-of-the-art process theory that explains how biological brains handle this dual problem. However, its applications to artificial intelligence have thus far been limited to extensions of existing model-based approaches. We present a formal abstraction of reinforcement learning algorithms that spans model-based, distributional, and model-free approaches. This abstraction seamlessly integrates active inference into the distributional reinforcement learning framework, making its performance advantages accessible without transition dynamics modeling.
BibTeX
@inproceedings{
akgul2026distributional,
title={Distributional Active Inference},
author={Abdullah Akg{\"u}l and Gulcin Baykal and Manuel Haussmann and Mustafa Mert {\c{C}}elikok and Melih Kandemir},
booktitle={Forty-third International Conference on Machine Learning},
year={2026},
url={https://openreview.net/forum?id=m3O4qYQ5hu}
}