Reproduce “Persistence, Exposure, and Credit Routing in Linear Temporal Predictive Coding” 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
Akira Tanase, Julian Gould, Jeffrey Seely
The authors analyze how linear tPC preserves a cue through a delay and exposes it later as temporal error. They characterize persistence and exposure, then test delayed recall with fixed spatial structure. After 10^4 steps, full temporal credit achieves median test MSE 0.00548 versus 0.97057 for standard tPC across 15 seeds; this is a controlled result, not a general solution.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.