Reproduce “NeuroHorizon: Long-Horizon Forecasting of Neural Population Activity with Hierarchical Memory” 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
Peng Jiang, Xiaoxuan Jia
NeuroHorizon forecasts future macaque motor-cortex activity from 500 ms of initial spikes, without further observations. It predicts population counts sequentially, retaining recent predictions in detail and compressing older ones; scheduled sampling trains it to use its own predictions. Across 10 sessions from three macaques, it led tested models at horizons of 100–1000 ms, though its 1000-ms lead over LFADS was narrow and scores declined later.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.