arXiv AI

Ballad: Bandit-Based LLM Routing for Automated Heuristic Discovery

Ballad is an online bandit algorithm that orchestrates a pool of large language models (LLMs) to improve automated heuristic discovery for NP‑hard combinatorial optimization problems. By selecting models adaptively, propagating credit through a DAG of heuristic genealogy, and discounting stale observations, Ballad matches or exceeds the best single LLM while reducing API costs by up to 65.6%. The study challenges the assumption that larger, more expensive models always yield better heuristics, showing that a well‑managed ensemble can outperform any individual model.

arXiv AI
Sep 2

Bandits in Prod: Hyperparameter Optimization at Inference Time

The paper introduces Online Hyperparameter Optimization (OHPO), framing it as an infinitely many‑armed bandit problem over mixed and conditional search spaces. It proposes the IMABO framework, which couples any bandit policy with any oracle for proposing new configurations, and presents IMOSS—a restart‑free anytime policy with provable regret bounds. Experiments show that IMABO, combined with practical oracles such as TPE, an incumbent‑mutation oracle, and a pretrained tabular foundation model, outperforms random search across a range of settings from classical ML models to LLM‑based agents.

By Louis Abraham, Tuan-Anh Nguyen, Nicolas Devatine
arXiv AI
Sep 25

Canopy: Exploiting Piecewise Smooth Tree Priors for Multi-Fidelity Bandits

CANOPY is a multi‑fidelity tree bandit algorithm that learns where a piecewise‑smooth prior holds instead of assuming global smoothness. It uses cheap random‑path probes to certify local aggregation bias and then focuses expensive leaf evaluations on cells where smoothness is violated. The method achieves provable fixed‑budget and regret guarantees that scale with the number of discontinuities, matching smooth‑tree rates when no violations exist and approaching structure‑blind search when violations are dense.

By Michael Jerge, Suman Jana
Hugging Face Trending Papers
Oct 5

Dynamic Minimax Regret Optimization for Robust LLM Post-Training

The paper introduces DUCB-OGD, an algorithm that couples a Discounted Upper‑Confidence‑Bound sampler with Online Gradient Descent to address dynamic minimax regret in robust large‑language‑model post‑training. It operates under instantaneous mini‑batch‑only bandit feedback, tracking worst‑source performance without re‑evaluating historical data. Experiments on fine‑tuning, preference optimization, and reinforcement learning demonstrate that DUCB‑OGD improves worst‑group robustness with negligible computational overhead.

arXiv Machine Learning
Oct 12

Accelerating Linear Bandits with Persistent Candidate Sets

The paper introduces Persistent Candidate LinUCB (PC‑LinUCB), a variant of the LinUCB algorithm for linear bandits that reduces per‑round score evaluations from ≥×K to ≥×(k+q) by maintaining a small set of active actions and sampling challengers from the remaining actions. It provides high‑probability regret bounds that include a standard learning term and an additive discovery cost of order ≥×((K-k)×log(K-k))/q, which does not grow with the horizon. Experiments demonstrate that PC‑LinUCB accelerates LinUCB while keeping regret comparable, and the persistent candidate approach can be applied beyond LinUCB.

By Bingkui Tong, Daria Korovaitceva, Marina Sheshukova, Sergey Samsonov, Junpei Komiyama