Research map / Label-efficient and robust learning / Recognition and long-tail
Neural Program Repair by Jointly Learning to Localize and Repair
Marko Vasic, Aditya Kanade, Petros Maniatis, David Bieber, Rishabh Singh
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 |
|---|---|---|---|
| Jointly predict bug presence, location, and repair instead of enumerating slot-wise repair predictions | Variable misuse repair on ETH-Py150 · ETH-Py150 test | Localization+Repair Accuracy: 55.8\% -> 65.7\% | Improved |
| Jointly predict bug presence, location, and repair instead of enumerating slot-wise repair predictions | Variable misuse repair on ETH-Py150 · ETH-Py150 test | Classification Accuracy: 71.1\% -> 82.4\% | Improved |
| Jointly predict bug presence, location, and repair instead of enumerating slot-wise repair predictions | Variable misuse repair on industrial snapshots · Industrial snapshot test | Localization+Repair Accuracy: 4.5\% -> 15.8\% | Improved |
| Jointly predict bug presence, location, and repair instead of enumerating slot-wise repair predictions | Variable misuse repair on industrial snapshots · Industrial snapshot test | Classification Accuracy: 47.2\% -> 66.7\% | Improved |
| Introduce an unrelated variable-misuse bug away from the prediction slot | Repair-only pointer model · AddBugAny vs. NoBugAny | Repair accuracy: 4.3 percentage points at τ=0.5 | Hurt |
| Introduce a nearby variable-misuse bug within two variable-use locations of the slot | Repair-only pointer model · AddBugNear vs. NoBugNear | Repair accuracy: 8 to 14.6 percentage points | Hurt |
| Use a pointer network over an RNN encoder instead of the syntax-only graph model | Variable misuse repair · MSR-VarMisuse seen test | Accuracy: 55.3\% -> 62.3\% | Improved |
Closest papers by meaning
- Counterexample Guided Program Repair Using Zero-Shot Learning and MaxSAT-based Fault Localization (AAAI 2025 · similarity 0.88)
- Graph-based, Self-Supervised Program Repair from Diagnostic Feedback (ICML 2020 · similarity 0.87)
- Less Is More: Adaptive Program Repair with Bug Localization and Preference Learning (AAAI 2025 · similarity 0.87)
- Logic Error Localization and Correction with Machine Learning (Student Abstract) (AAAI 2023 · similarity 0.86)
- Code Repair with LLMs gives an Exploration-Exploitation Tradeoff (NeurIPS 2024 · similarity 0.83)
- Guess & Sketch: Language Model Guided Transpilation (ICLR 2024 · 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.
