Research map / Label-efficient and robust learning / OOD and anomaly detection
On Distribution Shift in Learning-based Bug Detectors
Jingxuan He, Luca Beurer-Kellner, Martin Vechev
This paper tried a common change: Add a contrastive loss
Add contrastive learning · wrong-binop real-test · classification precision: 47.96 -> 52.30 Improved
What 1,317 accepted papers report for this change
- 1,253 improved
- 22 no clear effect
- 16 mixed
- 26 hurt
- Beyond Looking Up, Try Looking Around: Harmonizing Global Structure and Local Consistency in Optimal Transport for Short Text Clustering ICML 2026 · add contrastive learning · SearchSnippets · ACC: 56.65 -> 80.12Improved
- SleepLM: Natural-Language Intelligence for Human Sleep ICML 2026 · add contrastive objective to captioning (CoCa vs Cap) · sleep event IoU · IoU: 29.2 -> 27.4Hurt
- LIMSSR: LLM-Driven Sequence-to-Score Reasoning under Training-Time Incomplete Multimodal Observations ICML 2026 · add contrastive loss L_con · FS1000 average · Spearman correlation: 0.759 -> 0.789Improved
See this change in detailMatched by the wording of each paper's own result, read by AI from the full text. Papers report gains more often than failures, so this shows what gets published, not the chance a change works.
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 |
|---|---|---|---|
| Train first on synthetic bugs, then on imbalanced real bugs | CuBERT bug detectors · var-misuse real-test | classification precision: 3.43 -> 64.79 | Improved |
| Train first on synthetic bugs, then on imbalanced real bugs | CuBERT bug detectors · wrong-binop real-test | classification recall: 49.08 -> 43.99 | Mixed |
| Use real bugs rather than synthetic bugs in phase two (ablation) | Two-phase CuBERT training · var-misuse real-test | classification precision: 35.59 -> 64.79 | Improved |
| Use real bugs rather than synthetic bugs in phase two (ablation) | Two-phase CuBERT training · arg-swap real-test | classification precision: 44.19 -> 73.68 | Improved |
| Add a separate classification head (ablation) | CuBERT bug detectors · var-misuse real-test | classification precision: 49.45 -> 64.79 | Improved |
| Add task hierarchy (ablation) | CuBERT bug detectors · var-misuse real-test | classification precision: 51.82 -> 64.79 | Improved |
| Add focal loss (ablation) | CuBERT bug detectors · wrong-binop real-test | classification precision: 49.32 -> 52.30 | Improved |
| Add contrastive learning (ablation) | CuBERT bug detectors · wrong-binop real-test | classification precision: 47.96 -> 52.30 | Improved |
| Add focal loss (ablation) | CuBERT bug detectors · arg-swap real-test | classification precision: 73.68 -> 73.68 | No clear effect |
| Increase the non-buggy-to-buggy ratio in real-data training (ablation) | Two-phase CuBERT detector · var-misuse real-test | average precision: 7.40 -> 20.96 | Improved |
| Increase the non-buggy-to-buggy ratio in real-data training (ablation) | Two-phase CuBERT detector · wrong-binop real-test | average precision: 19.84 -> 39.85 | Improved |
| Increase the non-buggy-to-buggy ratio in real-data training (ablation) | Two-phase CuBERT detector · arg-swap real-test | average precision | No clear effect |
Closest papers by meaning
- Self-Supervised Bug Detection and Repair (NeurIPS 2021 · similarity 0.87)
- Neural Attribution for Semantic Bug-Localization in Student Programs (NeurIPS 2019 · similarity 0.86)
- HOPPITY: LEARNING GRAPH TRANSFORMATIONS TO DETECT AND FIX BUGS IN PROGRAMS (ICLR 2020 · similarity 0.84)
- Less Is More: Adaptive Program Repair with Bug Localization and Preference Learning (AAAI 2025 · similarity 0.83)
- Expecting The Unexpected: Towards Broad Out-Of-Distribution Detection (NeurIPS 2024 · similarity 0.81)
- Tolerant Algorithms for Learning with Arbitrary Covariate Shift (NeurIPS 2024 · similarity 0.80)
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.
