← back to paper
arxiv: 2506.05878 · 2 revisions
A projection-based framework for gradient-free and parallel learning