This insight led to MaxUCB, our new algorithm with strong theoretical guarantees and excellent empirical performance.

📍Meet @keggensperger.bsky.social at NeurIPS (San Diego) or @claireve.bsky.social and me at EurIPS (Copenhagen)!

💻 Code: github.com/amirbalef/CA...

(3/3)

“Max K-Armed Bandits should be perfect for HPO… but why are they never used?”

Because they are fundamentally harder than classic bandits and require proper assumptions. In the HPO setting, we identify exactly those assumptions and show how they enable efficient estimation of extremes.

(2/3)

Put CASH on Bandits: A Max K-Armed Problem for Automated Machine Learning The Combined Algorithm Selection and Hyperparameter optimization (CASH) is a challenging resource allocation problem in the field of AutoML. We propose MaxUCB, a max $k$-armed bandit method to trade o...

I am happy to share that our paper "Put CASH on Bandits: A Max K-Armed Problem for Automated Machine Learning" has been accepted at NeurIPS 2025!

Endless thanks to my amazing co-authors @claireve.bsky.social and @keggensperger.bsky.social

📄 Read it on arXiv: arxiv.org/abs/2505.05226

(1/3)