Reproduce “Simplified derivations for high-dimensional convex learning 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
David G. Clark, Haim Sompolinsky
Clark and Sompolinsky derive a non-replica cavity framework for high-dimensional random-data classification and kernel ridge regression. The method adds feature or datum variables and tracks how the opposite population responds through random couplings. It rederives established perceptron and manifold capacities, while deriving regression training and test errors consistent with prior analyses. These results rely on specified high-dimensional settings and assumptions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.