Research map / Large language models / Agents, code and math
Neural Attribution for Semantic Bug-Localization in Student Programs
Rahul Gupta, Aditya Kanade, Shirish Shevade
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 |
|---|---|---|---|
| use NBL instead of Tarantula-1 (one passing test) | bug localization · Student C programs, top-10 | programs localized: 964 -> 1164 | Improved |
| use NBL instead of Ochiai-1 (one passing test) | bug localization · Student C programs, top-1 | programs localized: 227 -> 294 | Improved |
| use NBL instead of Tarantula-*/Ochiai-* (all passing tests) | bug localization · Student C programs, top-1 and top-5 | programs localized: top-1: 294 vs 311/385; top-10: 1164 vs 1141/1151 | Mixed |
| use NBL instead of syntactic diff-based approach | bug localization · Student C programs, top-10 | programs localized: 623 -> 1164 | Improved |
| restrict k-means clustering (5 clusters) for baseline program search (ablation) | NBL · Student C programs | localization accuracy: <0.5% change in every metric | No clear effect |
| evaluate on both success and failure pairs instead of failure-only pairs | TCNN · Evaluation pairs | classification accuracy: 54.48% -> about 72% | Improved |
Closest papers by meaning
- On Distribution Shift in Learning-based Bug Detectors (ICML 2022 · similarity 0.86)
- Neural Program Repair by Jointly Learning to Localize and Repair (ICLR 2019 · similarity 0.86)
- Control Flow Graph Embedding Based on Multi-Instance Decomposition for Bug Localization (AAAI 2020 · similarity 0.85)
- Counterexample Guided Program Repair Using Zero-Shot Learning and MaxSAT-based Fault Localization (AAAI 2025 · similarity 0.84)
- CoRNStack: High-Quality Contrastive Data for Better Code Retrieval and Reranking (ICLR 2025 · similarity 0.82)
- Detecting AI-Generated Code Assignments Using Perplexity of Large Language Models (AAAI 2024 · similarity 0.82)
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.
