Reproduce “Bio-inspired Learning and Decision-Making with Probabilistic In-Memory Computing Hardware: Part 1” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Thomas Dalgaty, Eiji Kawasaki, Miguel de Prado, Devendra Vyas, Tommaso Salvatori
The paper proposes making uncertainty part of predictive coding by using noisy neural activity to sample hidden states and noisy synaptic weights to sample model parameters. In a shallow, linear network, Langevin updates combine prediction-error gradients, priors, and Gaussian noise. Weight variability represents epistemic uncertainty. The proposal has no experiments or hardware results; performance remains future work.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.