Active sensing with predictive coding and uncertainty minimization

Abdelrahman Sharafeldin; Nabil Imam; Hannah Choi

doi:10.1016/j.patter.2024.100983

Active sensing with predictive coding and uncertainty minimization

Patterns (N Y). 2024 May 3;5(6):100983. doi: 10.1016/j.patter.2024.100983. eCollection 2024 Jun 14.

Authors

Abdelrahman Sharafeldin^{1

2

3}, Nabil Imam^{1

2}, Hannah Choi^{1

3}

Affiliations

¹ ML@GT, Georgia Institute of Technology, Atlanta, GA 30332, USA.
² School of Computational Science and Engineering, Georgia Institute of Technology, Atlanta, GA 30332, USA.
³ School of Mathematics, Georgia Institute of Technology, Atlanta, GA 30332, USA.

Abstract

We present an end-to-end architecture for embodied exploration inspired by two biological computations: predictive coding and uncertainty minimization. The architecture can be applied to any exploration setting in a task-independent and intrinsically driven manner. We first demonstrate our approach in a maze navigation task and show that it can discover the underlying transition distributions and spatial features of the environment. Second, we apply our model to a more complex active vision task, whereby an agent actively samples its visual environment to gather information. We show that our model builds unsupervised representations through exploration that allow it to efficiently categorize visual scenes. We further show that using these representations for downstream classification leads to superior data efficiency and learning speed compared to other baselines while maintaining lower parameter complexity. Finally, the modular structure of our model facilitates interpretability, allowing us to probe its internal mechanisms and representations during exploration.

Keywords: active vision; embodied exploration; generative model; information maximization; intrinsic motivation; neuro-inspired AI; predictive coding; variational inference.