Beyond Similarity Scores: Evidence-Based Third-Party Library Detection for C/C++ Binaries
Published in ISSTA, 2026
🏆 ACM SIGSOFT Distinguished Paper Award (ISSTA 2026)
BLADE replaces similarity-score heuristics with multi-source, evidence-based verification powered by LLM-guided analysis, achieving 97.60% precision and 93.74% recall (F1 95.63%) and outperforming the best baseline by +41.83 F1 points. BLADE has been integrated into a commercial SCA product for compliance checks in regulated industries such as automotive.
Recommended citation: Chengyue Liu, et al. (2026). "Beyond Similarity Scores: Evidence-Based Third-Party Library Detection for C/C++ Binaries." Proceedings of the ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA 2026). ACM SIGSOFT Distinguished Paper Award.
Download Paper