Communication Dans Un Congrès Année : 2024

A Diffusion Model-Based Approach to Active Inference

Résumé

One of the current challenges in machine learning is the acquisition of generative models to enable decision making in complex, uncertain environments. In Bayesian methods like Active Inference, this requires the estimation of posterior distributions, which is often intractable. Variational Free Energy (VFE) provides a tractable approximation method for this estimation. This paper proposes a novel framework that leverages the expressive power of diffusion models to estimate free energy quickly and efficiently. Our approach, merging concepts from statistical physics, machine learning, and computational neuroscience, aims to improve the estimation of posterior distributions. We establish a theoretical link between active inference and diffusion models, emphasizing their shared strategies for uncertainty management, action selection, and model optimization. We propose using diffusion processes to approximate environmental generative models and employ scorebased guidance for action selection through expected free-energy minimization. Simulations in increasingly complex grid world environments demonstrate the framework's ability to handle partial observability and stochastic transitions, showing enhanced goaldirected behavior and uncertainty resolution compared to traditional methods. This suggests a path toward scalable active inference agents that could reason about uncertainty in real-world contexts, with possible applications to robotics, healthcare, and adaptive control systems.
Fichier principal
Vignette du fichier
paper_FMLDS-2.pdf (363) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04755289 , version 1 (27-10-2024)

Identifiants

  • HAL Id : hal-04755289 , version 1

Citer

Elias Aoun Durand, Mateus Joffily, Mehdi Khamassi. A Diffusion Model-Based Approach to Active Inference. 2024 IEEE International Conference on Future of Machine Learning and Data Science (FMLDS 2024), IEEE, Nov 2024, Sydney, Australia. ⟨hal-04755289⟩
0 Consultations
0 Téléchargements

Partager

More