AI Paper Atlas

Research map / Bandits and online learning / Online convex optimization and games

Robust Load Balancing with Machine Learned Advice

Sara Ahmadian, Hossein Esfandiari, Vahab Mirrokni, Binghui Peng

JMLR 2023 · Bandits and online learning › Online convex optimization and games · approach: Learning-augmented online algorithms · 8 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
increase mixing rate beta (more estimate error) from 0 to 1 (ablation)d-choice algorithm, Multinomial estimates · synthetic querieslimiting approximation ratio: 1.134 -> 1.729Hurt
increase mixing rate beta from 0 to 1 (ablation)d-choice algorithm, Gaussian estimates · synthetic querieslimiting approximation ratio: 1.01 -> 1.62Hurt
increase mixing rate beta from 0 to 1 (ablation)d-choice algorithm, Exponential estimates · synthetic querieslimiting approximation ratio: 1.03 -> 1.66Hurt
build the graph with randomized greedy instead of random or greedy allocation (real data) (ablation)d-choice algorithm · real query logs AS, EU, NAmax server loadNo clear effect
build the graph with randomized greedy instead of greedy or random allocation (synthetic) (ablation)d-choice algorithm · synthetic Multinomial, beta 0 to 1limiting approximation ratioMixed
run d-choice for more queries on real query logsd-choice algorithm · real query logs AS, EU, NAmax load ratio to lower bound: at most 1.5 -> 1Improved

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