When Labels Are Scarce: A Systematic Mapping of Label-Efficient Code Vulnerability Detection
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915904119373824 |
|---|---|
| author | Khalal, Noor Fettal, Chakib Labiod, Lazhar Nadif, Mohamed |
| author_facet | Khalal, Noor Fettal, Chakib Labiod, Lazhar Nadif, Mohamed |
| contents | Machine-learning-based code vulnerability detection (CVD) has progressed rapidly, from deep program representations to pretrained code models and LLM-centered pipelines. Yet dependable vulnerability labeling remains expensive, noisy, and uneven across projects, languages, and CWE types, motivating approaches that reduce reliance on human labeling. This survey maps these approaches, synthesizing five paradigm families and the mechanisms they use. It connects mechanisms to token, graph, hybrid, and knowledgebased representations, and consolidates evaluation and reporting axes that limit comparison (label-budget specification, compute/cost assumptions, leakage, and granularity mismatches). A Design Map and constraintfirst Decision Guide distill trade-offs and failure modes for practical method selection. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_00079 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | When Labels Are Scarce: A Systematic Mapping of Label-Efficient Code Vulnerability Detection Khalal, Noor Fettal, Chakib Labiod, Lazhar Nadif, Mohamed Cryptography and Security Software Engineering Machine-learning-based code vulnerability detection (CVD) has progressed rapidly, from deep program representations to pretrained code models and LLM-centered pipelines. Yet dependable vulnerability labeling remains expensive, noisy, and uneven across projects, languages, and CWE types, motivating approaches that reduce reliance on human labeling. This survey maps these approaches, synthesizing five paradigm families and the mechanisms they use. It connects mechanisms to token, graph, hybrid, and knowledgebased representations, and consolidates evaluation and reporting axes that limit comparison (label-budget specification, compute/cost assumptions, leakage, and granularity mismatches). A Design Map and constraintfirst Decision Guide distill trade-offs and failure modes for practical method selection. |
| title | When Labels Are Scarce: A Systematic Mapping of Label-Efficient Code Vulnerability Detection |
| topic | Cryptography and Security Software Engineering |
| url | https://arxiv.org/abs/2604.00079 |