📢 New in JMLR (w @rajivsambharya.bsky.social)! 🎉 Data-driven guarantees for classical & learned optimizers via sample bounds + PAC-Bayes theory. 📄 jmlr.org/papers/v26/2... 💻 github.com/stellatogrp/...

We learned the hyperparameters to accelerate algorithms over a family of problems. Turns out that we only need 10 training instances in each example and learn long steps for (prox) gd! Check out this work with @stellato.io paper: arxiv.org/pdf/2411.15717 code: github.com/stellatogrp/...