DeCoMa: Detecting and Purifying Code Dataset Watermarks through Dual Channel Code Abstraction

2025-04-09

AI Summary

Existing code watermarking techniques exhibit robustness against dilution and backdoor attacks but suffer from poor detectability and cleanability due to their high concealment. This paper proposes a dual-channel code abstraction framework enabling zero-shot watermark detection and silent purification without model training. Leveraging standardized template modeling and anomaly correlation analysis, the framework jointly exploits syntactic structures and semantic patterns to identify watermarks embedded at ultra-low rates (0.1%). Its key innovation is the first unsupervised outlier correlation detection mechanism under dual-channel constraints. Evaluated across 14 watermark types and 14 code-related tasks, the method achieves 100% stable recall, with detection speed accelerated by 31.5–130.9× over baselines. Crucially, post-purification model performance remains fully preserved, demonstrating both efficacy and practicality for real-world deployment.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Watermarking is a technique to help identify the source of data points, which can be used to help prevent the misuse of protected datasets. Existing methods on code watermarking, leveraging the idea from the backdoor research, embed stealthy triggers as watermarks.Despite their high resilience against dilution attacks and backdoor detections, the robustness has not been fully evaluated. To fill this gap, we propose DeCoMa, a dual-channel approach to Detect and purify Code dataset waterMarks.To overcome the high barrier created by the stealthy and hidden nature of code watermarks, DeCoMa leverages dual-channel constraints on code to generalize and map code samples into standardized templates. Subsequently, DeCoMa extracts hidden watermarks by identifying outlier associations between paired elements within the standardized templates. Finally, DeCoMa purifies the watermarked dataset by removing all samples containing the detected watermark, enabling the silent appropriation of protected code. We conduct extensive experiments to evaluate the effectiveness and efficiency of DeCoMa, covering 14 types of code watermarks and 3 representative intelligent code tasks (a total of 14 scenarios). Experimental results demonstrate that DeCoMa achieves a stable recall of 100% in 14 code watermark detection scenarios, significantly outperforming the baselines. Additionally, DeCoMa effectively attacks code watermarks with embedding rates as low as 0.1%, while maintaining comparable model performance after training on the purified dataset. Furthermore, as DeCoMa requires no model training for detection, it achieves substantially higher efficiency than all baselines, with a speedup ranging from 31.5 to 130.9X. The results call for more advanced watermarking techniques for code models, while DeCoMa can serve as a baseline for future evaluation.

Problem

• Detect and purify hidden watermarks in code datasets
• Evaluate robustness of existing code watermarking techniques
• Enable silent appropriation of protected code resources

Innovation

• Dual-channel constraints generalize code samples
• Extracts watermarks via outlier associations
• Purifies dataset by removing watermarked samples