Qi Pang (Carnegie Mellon University), Yuanyuan Yuan (HKUST), Shuai Wang (HKUST)

Secure multi-party computation (MPC) has recently become prominent as a concept to enable multiple parties to perform privacy-preserving machine learning without leaking sensitive data or details of pre-trained models to the other parties. Industry and the community have been actively developing and promoting high-quality MPC frameworks (e.g., based on TensorFlow and PyTorch) to enable the usage of MPC-hardened models, greatly easing the development cycle of integrating deep learning models with MPC primitives.

Despite the prosperous development and adoption of MPC frameworks, a principled and systematic understanding toward the correctness of those MPC frameworks does not yet exist. To fill this critical gap, this paper introduces MPCDiff, a differential testing framework to effectively uncover inputs that cause deviant outputs of MPC-hardened models and their plaintext versions. We further develop techniques to localize error-causing computation units in MPC-hardened models and automatically repair those defects.

We evaluate MPCDiff using real-world popular MPC frameworks for deep learning developed by Meta (Facebook), Alibaba Group, Cape Privacy, and OpenMined. MPCDiff successfully detected over one thousand inputs that result in largely deviant outputs. These deviation-triggering inputs are (visually) meaningful in comparison to regular inputs, indicating that our findings may cause great confusion in the daily usage of MPC frameworks. After localizing and repairing error-causing computation units, the robustness of MPC-hardened models can be notably enhanced without sacrificing accuracy and with negligible overhead.

View More Papers

Towards Real-time Voice Interaction Data Collection Monitoring and Ambient...

Tu Le (University of California, Irvine), Zixin Wang (Zhejiang University), Danny Yuxing Huang (New York University), Yaxing Yao (Virginia Tech), Yuan Tian (University of California, Los Angeles)

Read More

Automatic Policy Synthesis and Enforcement for Protecting Untrusted Deserialization

Quan Zhang (Tsinghua University), Yiwen Xu (Tsinghua University), Zijing Yin (Tsinghua University), Chijin Zhou (Tsinghua University), Yu Jiang (Tsinghua University)

Read More

WIP: Towards a Certifiably Robust Defense for Multi-label Classifiers...

Dennis Jacob, Chong Xiang, Prateek Mittal (Princeton University)

Read More

Compromising Industrial Processes using Web-Based Programmable Logic Controller Malware

Ryan Pickren (Georgia Institute of Technology), Tohid Shekari (Georgia Institute of Technology), Saman Zonouz (Georgia Institute of Technology), Raheem Beyah (Georgia Institute of Technology)

Read More