Reproduce “Near-Optimal Parallel $M$-Matrix Scaling” 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

Chih-Hung Liu, Aaron Sidford, Ta-Wei Tu

The authors present a randomized algorithm for scaling a symmetric positive-definite M-matrix so the rows of XAX sum approximately to 1. A nonnegative path initialization and infinity-norm centrality argument enable larger interior-point steps, reducing the iteration requirement to polylogarithmically many SDD solves. With high probability, it achieves residual at most ε and polylogarithmic parallel depth, with bounds depending on condition number and accuracy.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.