Reproduce “Optimizing Rank for High-Fidelity Implicit Neural Representations” 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
Julian McGinnis, Florian A. Hölzl, Suprosanna Shit, Florentin Bieder, Paul Friedrich, Mark Mühlau, Bjoern Menze, Daniel Rueckert and 1 more
Paper shows optimization can help a ReLU MLP fit fine image detail. Muon orthogonalizes hidden-layer weight updates while Adam handles other parameters, preserving higher stable rank without changing the activation or input encoding. On 24 Kodak images, mean PSNR rises from 22.83 ± 2.85 to 30.04 ± 3.59 dB, narrowing but not closing the gap to SIREN and FINER.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.