Contributed Talk Sessions | Poster Sessions | All Posters | Search Papers

Poster Session A: Tuesday, August 12, 1:30 – 4:30 pm, de Brug & E‑Hall

Predictive Coding algorithms induce brain-like responses in Artificial Neural Networks

Dirk C. Gütlin¹, Ryszard Auksztulewicz²; ¹Freie Universität Berlin, ²Maastricht University

Presenter: Dirk C. Gütlin

This study investigates whether predictive coding (PC) inspired deep neural networks can serve as biologically plausible models of the brain. We compared two PC-inspired training objectives - a predictive and a contrastive approach - to a supervised baseline, using a simple recurrent neural network (RNN) architecture. Our results show that, compared to Supervised or Untrained models, the PC-inspired models exhibited more key signatures of PC. This includes mismatch responses (MMR), formation of prior expectations, and learning of semantic representations. These findings indicate that PC-inspired models can capture important computational principles of predictive processing in the brain, and serve as a promising foundation for building biologically plausible artificial neural networks.

Topic Area: Predictive Processing & Cognitive Control

Extended Abstract: Full Text PDF