Professor Minsoo Rhu’s research lab has been selected for the 2024 SW Star Lab Project under the Information and Communication Broadcasting Technology Development Program
<Professor Minsoo Rhu>
Our department’s Professor Minsoo Rhu’s research lab has been selected for the 2024 SW Star Lab Project under the Information and Communication Broadcasting Technology Development Program, overseen by the Ministry of Science and ICT and the Institute for Information and Communications Technology Planning & Evaluation (IITP).
The “SW Star Lab” project aims to secure world-class original technology in the software field and train master’s and doctoral-level talents. The selected research lab will receive a total of 1.5 billion KRW (about 200 million KRW annually) over eight years.
Professor Rhu’s lab proposed a project titled “High-performance Privacy-preserving Machine Learning System Software for Cloud Computing” in the ‘Cloud/Computing’ field. The proposed research aims to develop an AI learning pipeline that can protect personal information. Privacy-protecting AI learning involves three stages: data processing and generation to create learning data, data storage and analysis, and AI model learning. The goal is to develop a next-generation cloud system that can perform these stages at high performance.
Big tech companies are making significant efforts to improve the quality of various AI services by collecting large amounts of user data to train AI models. However, this process involves serious concerns about user privacy protection. Sensitive personal information may be leaked during data transmission, storage, analysis at data centers, or during inference services where the trained models might expose the sensitive personal information . The proposed research aims to address these AI privacy issues and contribute to the establishment of safe AI technologies.
This research is expected to significantly reduce the costs of model training and inference for privacy-protecting AI services. This is crucial in which enhanced privacy regulations increase the computational complexity and resource requirements for AI learning, making the technologies economically challenging. Consequently, it is anticipated that this project will help reduce carbon emissions and secure an edge in commercializing the technology.
