Leaded by Prof.Xiaojun Chen, we are undering Data Security Laboratory of the Institute of Information Engineering (IIE), Chinese Academy of Sciences (CAS).

Established with a passion for scientific exploration and innovation, our laboratory is a hub of cutting-edge research and discovery. Our laboratory is primarily engaged in researching Artificial Intelligence Security, covering areas such as Privacy Computing, Federated Learning, Backdoor Attacks, Model Watermarking, and AI Security. In the realm of privacy computing, we strive to safeguard data privacy while enabling efficient computational processes. Federated learning offers a collaborative approach to model training without sacrificing data privacy. Backdoor attacks pose a serious threat, and we are constantly on the lookout for effective countermeasures. Model watermarking helps protect intellectual property and ensure the integrity of models. Additionally, we are actively investigating the security of large models to meet the emerging challenges in this rapidly evolving field.

Our team consists of four teachers, thirteen doctoral candidates, two master candidates and one visiting student. With a diverse group of talented individuals, our laboratory is a vibrant hub of research and innovation. We look forward to sharing our findings and collaborating with others in the pursuit of a more secure artificial intelligence future.

🔥 Latest News

• 2024.08 🎉🎉 Client Relevance-Aware Adaptive Aggregation for Personalized Federated Learning has been accepted by ICONIP2024!
• 2024.08 🎉🎉 Progtuning: Progressive Fine-tuning Framework for Transformer-based Language Models has been accepted by ICONIP2024!
• 2024.08 🎉🎉 CoFD: Contribution-based Federated Knowledge Aggregation Scheme for Federated Distillation has been accepted by ICONIP2024!
• 2024.08 🎉🎉 FLIGHT: Lightweight and Backdoor-Resistant Secure Aggregation has been accepted by ACSAC2024!
• 2024.07 🎉🎉 Privacy-preserving Desion Tree has been accepted by SecureComm 2024!
• 2024.07 🎉🎉 CipherDM: Secure Three-Party Inference for Diffusion Model Sampling has been accepted by ECCV2024!
• 2024.03 🎉🎉 DualCOS: Query-Efficient Data-Free Model Stealing with Dual Clone Networks and Optimal Samples has been accepted by ICME2024!
• 2024.03 🎉🎉 STMS: An Out-Of-Distribution Model Stealing Method Based on Causality has been accepted by IJCNN2024!
• ……

🏆 Selected Awards

• First prize in Privacy-Preserving Computing Hackathon, 2022

💻 Programs

• Key research and development projects of Shenzhen.
• Projects of Beijing Municipal Science and Technology Commission.
• Federated learning and privacy computing platform: Bonfire-IIE

💂‍♂️ Lab Members

📃 For more info

More info about our laboratory can be found in the Lab_Page. If you want to contact us, please send email to zhaoxin@iie.ac.cn.