Reproduce “Poisson Hyperplane Processes with Rectified Linear Units” 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
Shufei Ge, Shijia Wang, Lloyd Elliott
The authors represent hidden ReLU units as hyperplanes generated by a Poisson hyperplane process, assign Gaussian priors to output contributions, and fit the model using Bayesian inference and annealed sequential Monte Carlo. The number of planes is prespecified. In simulations, coordinate-domain decomposition had similar test RMSE and 4.69% of baseline per-job running time; this finding is limited to that setup.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.