Research map / Reinforcement learning / Visual and representation RL
Implementing Troubleshooting with Batch Repair
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 |
|---|---|---|---|
| Repair batches chosen by Pes.(2) instead of single most-likely component (1-HP) | 1-HP · 74181, overhead 25 | average repair cost: 239 -> 110 | Helped |
| Pes.(2) instead of BD-batch | BD-batch · c499, overhead 25 | average repair cost: 161 -> 69 | Helped |
| Pes.(2) instead of 1-HP | 1-HP · c880, overhead 25 | average repair cost: 287 -> 205 | Helped |
| Pes.(2) instead of 1-HP | 1-HP · 74182, overhead 10 | average repair cost: 82 -> 56 | Helped |
| Pes.(2) instead of 1-HP | 1-HP · 74182, overhead 25 | average repair cost: 164 -> 73 | Helped |
| Increase repair overhead from 10 to 25 (advantage of batch repair over 1-HP) | Pes.(2) vs 1-HP · 74283 | average repair cost gap: 41 -> 122 | Helped |
| Trivial BD-batch (repair whole best diagnosis) versus 1-HP | 1-HP · c432, overhead 25 | average repair cost: 120 -> 161 | Hurt |
| BD-batch versus 1-HP | 1-HP · 74182, overhead 25 | average repair cost: 164 -> 103 | Helped |
| Increase k from 1 to 2 in optimistic union-based search (ablation) | Opt.(k) · 74182, overhead 25 | average repair cost: 97 -> 82 | Helped |
| Increase k from 1 to 2 in pessimistic union-based search (ablation) | Pes.(k) · 74283, overhead 25 | average repair cost: 128 -> 91 | Helped |
| Increase k from 1 to 2 in pessimistic union-based search (ablation) | Pes.(k) · c880, overhead 25 | average repair cost: 225 -> 205 | Helped |
| Powerset-based search instead of union-based search (ablation) | Union-based search · all systems | average repair cost | Hurt |
Closest papers by meaning
- Counterexample Guided Program Repair Using Zero-Shot Learning and MaxSAT-based Fault Localization (AAAI 2025 · similarity 0.80)
- Less Is More: Adaptive Program Repair with Bug Localization and Preference Learning (AAAI 2025 · similarity 0.80)
- Is Self-Repair a Silver Bullet for Code Generation? (ICLR 2024 · similarity 0.78)
- Combining Cost Constrained Runtime Monitors for AI Safety (NeurIPS 2025 · similarity 0.78)
- PatchBench: Measuring Collateral Damage in Activation Patching (NeurIPS 2026 · similarity 0.77)
- Precise Debugging Benchmark: Is Your Model Debugging or Regenerating? (NeurIPS 2026 · similarity 0.76)
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.