Privacy-Preserving Machine Learning

· 1 min read

We study algorithmic and systems techniques for private AI pipelines in both centralized and federated settings.

Representative References

  1. Pencil: Private and Extensible Collaborative Learning without Non-Colluding Assumption . NDSS24 , 2024.
  2. CoGNN: Towards Secure and Efficient Collaborative Graph Learning . CCS24 , 2024.
  3. RingSG: Optimal Secure Vertex-Centric Computation for Collaborative Graph Processing . CCS25 , 2025.
  4. Defending Against Data Reconstruction Attacks in Federated Learning: An Information Theory Approach . Sec24 , 2024.
Zhuotao Liu
Authors
Zhuotao Liu
Associate Professor