Safe Artificial Intelligence

안전한 인공지능

관련기사 바로가기

Our research is focused on addressing the interconnected challenges of privacy, fairness, and security to promote the safe use of artificial intelligence (AI) algorithms in real-world systems. Our goal is to develop innovative solutions that enable privacy-preserving, fairness-aware, and security-enhanced machine learning. To achieve this goal, we are pursuing two key problem thrusts:
(i) We are developing comprehensive approaches for reliable AI while considering security and privacy threats.
(ii) We are addressing the need for realistic threat models and evaluation for security-aware algorithms.
We are committed to advancing the field of artificial intelligence in a responsible and ethical manner. We believe that these three pillars are critical for building AI systems that are safe and beneficial for individuals, industry, and society.

Major research field

Security-enhanced Machine Learning, Privacy-preserving Machine Learning, Fairness-aware Machine Learning

Desired field of research

Privacy evaluation for synthetic data, Practical attack models for artificial intelligence

Research Keywords and Topics

Safe Artificial Intelligence
- Security-enhanced Machine Learning
- Privacy-preserving Machine Learning
- Fairness-aware Machine Learning

Research Publications

Saerom Park, Sungmin Kim* and Yeon-sup Lim, “Fairness Audit of Machine Learning Models with Confidential Computing”, The ACM Web Conference (WWW) 2022, Apr. 25-29, 2022
Saerom Park and Jaewook Lee*. “Stability Analysis of Denoising Autoencoder based on Dynamical Projection System”, IEEE Transactions on Knowledge Data Engineering, Aug. 2021
Sungyoon Lee, Jaewook Lee and Saerom Park*, “Lipschitz-Certifiable Training with a Tight Outer Bound”, 34th Conference on Neural Information Processing Systems (NeurIPS 2020), Dec. 7-12, 2020