SMS: Self-supervised Model Seeding for Verification of Machine Unlearning

2025-09-29

AI Summary

Existing data unlearning verification methods rely on backdoor samples, making it difficult to assess a model’s actual removal of user-specific real-world data. To address this, we propose the Self-supervised Model Seeding (SMS) framework—the first approach to establish a verifiable, tripartite binding among user-specific seeds, original training samples, and the model itself, enabling rigorous unlearning verification on authentic data. SMS jointly optimizes a self-supervised seeding task with the primary learning objective, implicitly embedding user-unique seeds into the model’s latent space. This design preserves model utility and seed stealth while enabling trustworthy, post-deletion auditing of real-data removal. Experiments demonstrate that SMS significantly improves both the accuracy and practicality of unlearning verification on genuine samples—without compromising model performance—thereby overcoming the fundamental deployment limitations inherent in backdoor-based verification.

Technology Category

• Machine Learning: Unsupervised & Self-Supervised Learning
• Search and Optimization: Learning to Search
• Data Mining & Knowledge Management: Representing, Reasoning, and Using Provenance, Trust

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Security and Privacy: Data transparency and provenance

Abstract

Many machine unlearning methods have been proposed recently to uphold users' right to be forgotten. However, offering users verification of their data removal post-unlearning is an important yet under-explored problem. Current verifications typically rely on backdooring, i.e., adding backdoored samples to influence model performance. Nevertheless, the backdoor methods can merely establish a connection between backdoored samples and models but fail to connect the backdoor with genuine samples. Thus, the backdoor removal can only confirm the unlearning of backdoored samples, not users' genuine samples, as genuine samples are independent of backdoored ones. In this paper, we propose a Self-supervised Model Seeding (SMS) scheme to provide unlearning verification for genuine samples. Unlike backdooring, SMS links user-specific seeds (such as users' unique indices), original samples, and models, thereby facilitating the verification of unlearning genuine samples. However, implementing SMS for unlearning verification presents two significant challenges. First, embedding the seeds into the service model while keeping them secret from the server requires a sophisticated approach. We address this by employing a self-supervised model seeding task, which learns the entire sample, including the seeds, into the model's latent space. Second, maintaining the utility of the original service model while ensuring the seeding effect requires a delicate balance. We design a joint-training structure that optimizes both the self-supervised model seeding task and the primary service task simultaneously on the model, thereby maintaining model utility while achieving effective model seeding. The effectiveness of the proposed SMS scheme is evaluated through extensive experiments, which demonstrate that SMS provides effective verification for genuine sample unlearning, addressing existing limitations.

Problem

• Verifying genuine sample removal in machine unlearning systems
• Overcoming backdoor limitations for authentic data deletion confirmation
• Embedding user-specific seeds while maintaining model utility

Innovation

• Self-supervised model seeding links user seeds to genuine samples
• Joint-training balances model utility with seeding effectiveness
• Latent space embedding enables verification without server seed exposure