spnn-shortest-path: stochastic spiking networks for shortest-path computation

作者：Gary R. Engler · 发表于：Zenodo (CERN European Organization for Nuclear Research) · 年份：2026 · DOI：10.5281/zenodo.22035801 · 研究领域：Computer science、Algorithm、Theoretical computer science、Artificial intelligence、Data mining

Reference implementation, data, and figures for the paper on stochastic spiking neural networks for shortest-path computation. A weighted graph is compiled into winner-take-all neuron clusters whose Boltzmann energy minima coincide with shortest paths, and run as a discrete-time Boltzmann sampler at O(|E|) cost per step. Dual-licensed: the code is under the MIT License; the data (the contents of instances/ and results/) is under the Creative Commons Attribution 4.0 International License (CC BY 4.0). See the LICENSE file in the archive for the full terms.