Research Interests
My research interests are at the intersection of security, privacy, AI. Specifically, most of our research aims at enhancing the trustworthiness, safety, and efficiency of AI systems, as well as investigating the application of AI in enhancing security measures. Recently, my primary research interests are as follows.
1. AI Security and Privacy.
The emerging deep learning technology has been widely commercialized in our daily life. However, it also introduces new security and privacy threats, which could bring disastrous consequences in critical scenarios. Our goal is to investigate the security vulnerabilities as well as defense solutions for various deep learning systems, such as computer vision, natural language processing, reinforcement learning, generative AI, etc. Our recent work has focused on Large Language Models (LLMs), Large Vision Models (LVMs), and Multi-modal Large Models (MLMs). We are interested in the following security problems.
• Prompt Engineering and Safety
• Model Security and Privacy
• Training Dataset Privacy
• Intellectual Property (IP) Protection
• Large Models Memorization
2. AI Safety.
Artificial Intelligence (AI) has experienced remarkable growth and innovation, significantly impacting various sectors. However, this rapid development has also led to concerns about potential misuse, notably in the emergence of AI safety challenges. A key issue in this domain is the creation of DeepFakes, which use advanced deep learning techniques to produce highly convincing fake media content. This technology poses significant risks, such as spreading misinformation, identity theft, and manipulating public opinion. As DeepFake technology becomes more sophisticated, identifying and countering their effects is increasingly crucial. In fighting against DeepFakes, we are interested in the following security problems.
• DeepFake Passive Detection
• DeepFake Proactive Defense
• DeepFake Evasion Attack
• AIGC for DeepFake Creation
3. Efficient AI.
While AI models are improving very rapidly in performance, their computing and storage resource consumption is also rapidly increasing. As a response to this, efficient AI becomes a critical fields where researchers are working on creating more efficient, low-resource-demand and energy-saving AI models. Efficient AI will achieve the above goals by improving optimization algorithms, compressing models, and designing specialized hardware. The ultimate goal of research is to create powerful and environmentally friendly AI and make AI available everywhere. We are interested in the following problems.
• Model Compression
• Model Acceleration
