AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

Self-Supervised Bug Detection and Repair

Miltiadis Allamanis, Henry Richard Jackson-Flux, Marc Brockschmidt

NeurIPS 2021 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Self-supervision in supervised pipelines · 36 citations (OpenAlex)

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 changedWhereMeasuredEffect
Replace random bug insertion with a learned adversarial bug selector (ablation)PyBugLab · PyPIBugsJoint accuracy: 21.2 -> 24.2Improved
Replace random bug insertion with a learned adversarial bug selector (ablation)PyBugLab · PyPIBugsJoint accuracy: 20.6 -> 24.0Improved
Add semantics-preserving code augmentation (ablation)PyBugLab GNN · PyPIBugsJoint accuracy: 24.2 -> 26.4Improved
Add semantics-preserving code augmentation (ablation)PyBugLab GREAT · PyPIBugsJoint accuracy: 24.0 -> 23.2Hurt
Add semantics-preserving code augmentation (ablation)PyBugLab GNN · RandomBugsJoint accuracy: 69.6 -> 70.3Improved
Add semantics-preserving code augmentation (ablation)PyBugLab GREAT · RandomBugsJoint accuracy: 64.0 -> 65.3Improved
Use a learned bug selector rather than random bug insertion (ablation)PyBugLab · RandomBugsJoint accuracy: 69.4 -> 69.6No clear effect
Use a learned bug selector rather than random bug insertion (ablation)PyBugLab · RandomBugsJoint accuracy: 63.9 -> 64.0No clear effect
Use GNN instead of GREAT for code representationsPyBugLab+Aug · PyPIBugsJoint accuracy: 23.2 -> 26.4Mixed
Use PyBugLab rather than CuBERTBug detection · PyPIBugs-WrongOpF1: 0.378 -> 0.746Improved
Use PyBugLab rather than CuBERTBug detection · PyPIBugs-VarMisuseF1: 0.493 -> 0.787Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy