Reproduce “Thresholds of descending algorithms in inference problems” 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
Stefano Sarao Mannelli, Lenka Zdeborova
The authors analyze recovery in a spiked matrix–tensor model by comparing AMP, Langevin sampling, and gradient flow through its non-convex landscape. Lower tensor noise can delay Langevin’s overlap, while gradient flow detects the signal before spurious minima disappear. For p = 3 and Δ₃ = 1, its predicted threshold is 1/Δ₂ = 2, a model-specific result, not a general guarantee.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.