Research map / Multi-agent systems and games / Games and equilibria
Curriculum learning for multilevel budgeted combinatorial problems
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Train with a bottom-up curriculum over increasing budgets, using exact or heuristic afterstate targets from smaller-budget experts | MultiL-Cur · MCN, D^(1) test sets | optimality gap (%): 0.51 vs 3.34 vs 14.53 | Helped |
| Use MultiL-Cur instead of DA-AD | MCN · MCN, |V|=40 | optimality gap (%): 7.6 -> 5.0 | Helped |
| Use MultiL-Cur instead of DA-AD | MCN · MCN, |V|=60 | optimality gap (%): 7.3 -> 4.4 | Helped |
| Use MultiL-Cur instead of DA-AD | MCN · MCN, |V|=80 | optimality gap (%): 3.8 -> 2.7 | Helped |
| Use MultiL-Cur instead of DA-AD | MCN · MCN, |V|=100 | optimality gap (%): 2.7 -> 49.6 | Hurt |
| Increase training episodes from 960 000 to 7 680 000 (ablation) | MultiL-DQN · MCN, Vaccination | optimality gap (%): 29.8 -> 6.7 | Helped |
| Increase training episodes from 960 000 to 7 680 000 (ablation) | MultiL-DQN · MCN, Attack | optimality gap (%): 35.8 -> 21.2 | Helped |
| Increase training episodes from 960 000 to 7 680 000 (ablation) | MultiL-DQN · MCN, Protection | optimality gap (%): 28.8 -> 4.01 | Helped |
| Use MultiL-Cur instead of MultiL-MC and MultiL-DQN | MCN · MCN, D^(1) test sets | approximation ratio: 1.006 vs 1.036 vs 1.158 | Helped |
| Evaluate MultiL-Cur on weighted and directed MCN variants | MultiL-Cur · MCN_dir and MCN_w, |V|=20–100 | optimality gap (%) | Mixed |
Closest papers by meaning
- GOAL: A Generalist Combinatorial Optimization Agent Learner (ICLR 2025 · similarity 0.89)
- Learning Combinatorial Optimization Algorithms over Graphs (NeurIPS 2017 · similarity 0.89)
- Optimizing Solution-Samplers for Combinatorial Problems: The Landscape of Policy-Gradient Method (NeurIPS 2023 · similarity 0.88)
- Structured Reinforcement Learning for Combinatorial Decision-Making (NeurIPS 2025 · similarity 0.87)
- PolyNet: Learning Diverse Solution Strategies for Neural Combinatorial Optimization (ICLR 2025 · similarity 0.87)
- Are Graph Neural Networks Optimal Approximation Algorithms? (NeurIPS 2024 · similarity 0.86)
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.