Research map / Learning theory and algorithms / Prediction and decision losses
Applied Online Algorithms with Heterogeneous Predictors
Jessica Maghakian, Russell Lee, Mohammad Hajiesmaili, Jian Li, Ramesh Sitaraman, Zhenhua Liu
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 changed | Where | Measured | Effect |
|---|---|---|---|
| classification meta-algorithm selecting among DBO, parSBO, inDBO instead of a single predictor-based algorithm | bandwidth cost minimization with two parameter predictors · Akamai production traffic, 180+ metro-areas | worst case normalized cost: 1.236 vs 2.817 (parSBO-1) and 6.107 (parSBO-2) | Improved |
| classification meta-algorithm selecting among DBO, parSBO, inDBO | bandwidth cost minimization · Akamai production traffic, 180+ metro-areas | average case normalized cost: 1.022 vs 1.014 for parSBO-1 (meta slightly worse than parSBO-1) | No clear effect |
| increasing trust parameter λ in WLW (less reliance on historical data) (ablation) | WLW (Wang et al. 2020) multi-prediction algorithm · Akamai production traffic, 180+ metro-areas | worst case normalized cost: worst case rises from 1.630 (λ=0.25) to 1.730 (λ=0.75) | Hurt |
| adding input predictions (inDBO-style, robustifying parameter predictor) to parameter prediction via RoBO-2 with low λ (ablation) | RoBO on COVID-19 Milan lockdown data · Milan COVID-19 lockdown demand, first month | normalized cost: RoBO-2 beats parSBO and inDBO alone at low λ | Improved |
| partial trust (λ≥0.3) in parameter prediction in RoBO-1 (ablation) | RoBO-1 on COVID-19 lockdown data · Milan COVID-19 lockdown demand, first month | normalized cost | Mixed |
| relying fully on parameter predictions (parSBO) after the shift stabilizes (ablation) | parSBO on April-May COVID-19 data · Milan COVID-19 lockdown demand, April and May | normalized cost | Improved |
Closest papers by meaning
- Learning-Augmented Algorithms with Explicit Predictors (NeurIPS 2024 · similarity 0.86)
- Online Algorithms with Uncertainty-Quantified Predictions (ICML 2024 · similarity 0.85)
- FedRSPO+: A Heterogeneity-aware Algorithm for Decision-focused Federated Learning (NeurIPS 2026 · similarity 0.84)
- Coded Computing for Resilient Distributed Computing: A Learning-Theoretic Framework (NeurIPS 2024 · similarity 0.84)
- ATA: Adaptive Task Allocation for Efficient Resource Management in Distributed Machine Learning (ICML 2025 · similarity 0.83)
- The Stability of Online Algorithms in Performative Prediction (NeurIPS 2026 · similarity 0.83)
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.
