Jing Shang (Beijing Jiaotong University), Jian Wang (Beijing Jiaotong University), Kailun Wang (Beijing Jiaotong University), Jiqiang Liu (Beijing Jiaotong University), Nan Jiang (Beijing University of Technology), Md Armanuzzaman (Northeastern University), Ziming Zhao (Northeastern University)

Model pruning is a technique for compressing deep learning models, and using an iterative way to prune the model can achieve better compression effects with lower utility loss. However, our analysis reveals that iterative pruning significantly increases model memorization, making the pruned models more vulnerable to membership inference attacks (MIAs). Unfortunately, the vast majority of existing defenses against MIAs are designed for original and unpruned models. In this paper, we propose a new framework WeMem to weaken memorization in the iterative pruning process. Specifically, our analysis identifies two important factors that increase memorization in iterative pruning, namely data reuse and inherent memorability. We consider the individual and combined impacts of both factors, forming three scenarios that lead to increased memorization in iteratively pruned models. We design three defense primitives based on these factors' characteristics. By combining these primitives, we propose methods tailored to each scenario to weaken memorization effectively. Comprehensive experiments under ten adaptive MIAs demonstrate the effectiveness of the proposed defenses. Moreover, our defenses outperform five existing defenses in terms of privacy-utility tradeoff and efficiency. Additionally, we enhance the proposed defenses to automatically adjust settings for optimal defense, improving their practicability.

View More Papers

Automatic Library Fuzzing through API Relation Evolvement

Jiayi Lin (The University of Hong Kong), Qingyu Zhang (The University of Hong Kong), Junzhe Li (The University of Hong Kong), Chenxin Sun (The University of Hong Kong), Hao Zhou (The Hong Kong Polytechnic University), Changhua Luo (The University of Hong Kong), Chenxiong Qian (The University of Hong Kong)

Read More

Duumviri: Detecting Trackers and Mixed Trackers with a Breakage...

He Shuang (University of Toronto), Lianying Zhao (Carleton University and University of Toronto), David Lie (University of Toronto)

Read More

Towards Establishing a Systematic Security Framework for Next Generation...

Tolga O. Atalay (A2 Labs LLC), Tianyuan Yu (UCLA), Lixia Zhang (UCLA), Angelos Stavrou (A2 Labs LLC)

Read More

Deanonymizing Device Identities via Side-channel Attacks in Exclusive-use IoTs...

Christopher Ellis (The Ohio State University), Yue Zhang (Drexel University), Mohit Kumar Jangid (The Ohio State University), Shixuan Zhao (The Ohio State University), Zhiqiang Lin (The Ohio State University)

Read More