Reproduce “Optimal Injectivity of the SparseStack” 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
Lorenzo Beretta, Simone Bombari, Margalit Glasgow, Raphael Arkady Meyer, Kijun Shin, David P. Woodruff
The authors analyze SparseStack sketches for preserving vectors in a fixed subspace without requiring control of stretching. Two independent halves handle spread-out vectors and isolate coordinates carrying concentrated mass. With linear sketch dimension, row sparsity O(sqrt(log r)) yields a constant lower bound on squared lengths with probability at least 19/20; the matching lower bound holds only under stated overlap conditions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.