AI Paper Atlas

Research map / Bandits and online learning / Contextual bandits and Thompson sampling

Online Multi-LLM Selection via Contextual Bandits Under Unstructured Context Evolution

Manhin Poon, Xiangxiang Dai, Xutong Liu, Fang Kong, John C.S. Lui, Jinhang Zuo

AAAI 2026 · Bandits and online learning › Contextual bandits and Thompson sampling · approach: Oracle-efficient contextual bandits · 19 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use multi-step contextual bandit selection (Greedy LinUCB) instead of single-step routersmulti-LLM routing · Average of 4 benchmarksaccuracy (%): MixLLM 64.25 -> Greedy LinUCB 71.99Improved
add positionally-aware knapsack heuristic (PAKH) instead of budget-aware greedy (ablation)B.A. LinUCB · Average of 4 benchmarksaccuracy (%): 69.17 -> 74.84Improved
add PAKH instead of Greedy LinUCB on average steps (ablation)LinUCB variants · Average of 4 benchmarksaverage steps: 2.52 -> 1.58Improved
add a budget constraint to Greedy LinUCB (ablation)Greedy LinUCB · Average of 4 benchmarksaccuracy (%) and steps: accuracy 71.99 -> 69.17; steps 2.52 -> 2.04; cost 1.51E-03 -> 2.63E-04Mixed
use B.A. LinUCB instead of MixLLM / MetaLLMmulti-LLM routing · Average of 4 benchmarksaccuracy (%) and cost: MixLLM 64.25 -> 69.17; cost 2.27E-03 -> 2.63E-04Improved
use PAKH instead of B.A. LinUCB at tight budgets (below $0.001) (ablation)budget-constrained selection · budget sensitivityaccuracyHurt
increase per-query budget (ablation)PAKH vs B.A. LinUCB · budget sensitivityaccuracyMixed
use adaptive LLM choice per step instead of retrying the initial router choiceMetaLLM, MixLLM · AIMEaccuracy (%): MetaLLM 20.83 -> Greedy LinUCB 54.16Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy