Ms Ronghui Mu

My research focuses on evaluating and improving the robustness of deep neural networks (DNNs), particularly in the context of safety-critical applications. This includes work on adversarial attack and defense, formal robustness verification, and systematic safety testing of DNNs.

I have conducted robustness analysis across a range of AI systems, including:

• Image classifiers
• Video recognition models
• 3D point cloud networks
• Reinforcement learning agents
• Large language models (LLMs)

My long-term vision is to build safe, reliable, and trustworthy deep learning systems that are robust under real-world uncertainties and malicious perturbations.

I am currently looking for PhD students interested in the following topics:

• Safe, Secure, and Explainable AI
• Adversarial Machine Learning and Robustness
• Probabilistic Verification and Formal Methods in AI
• Reinforcement Learning and its application
• NLP, Computer Vision, and generative AI (LLMs, VLMs)

Candidates with strong motivation, a solid background in machine learning, and an interest in AI safety are encouraged to get in touch.