Reproduce “Stochastic inference with deterministic spiking neurons” 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

Mihai A. Petrovici, Johannes Bill, Ilja Bytschok, Johannes Schemmel, Karlheinz Meier

The paper shows how deterministic leaky integrate-and-fire (LIF) neurons can support stochastic inference when noisy spiking input makes their firing statistics approximately probabilistic. A first-passage-time analysis derives a logistic activation function, followed by recurrent-network simulations. In a five-neuron model, sampled states approximate Boltzmann-machine targets, but KL divergence remains above zero at long times.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.