Research

Research

Our work spans computer vision and graphics, natural language and large language models, trustworthy data, and applied AI, often at their intersections.

Areas

Computer Vision & Graphics
From low-level vision and image restoration to 3D reconstruction, generative imaging, and medical image computing, we build models that perceive and synthesize the visual world.

Topics: low-level vision, 3D reconstruction, layout diffusion, medical image computing

Natural Language Processing & Large Language Models
We advance instruction tuning, code editing with LLMs, low-resource multilingual NLP, and the use of language models to assist software engineering.

Topics: instruction tuning, code editing, multilingual NLP, LLMs for software engineering

Trustworthy Data: Privacy, Security & Analytics
We study secure and efficient machine-learning lifecycles with trusted hardware, video authentication, watermarking, and privacy-preserving analytics.

Topics: trusted hardware, data privacy, authentication, time-series analytics

Applied & Industrial AI
We bring AI to real-world domains: generative AI for music, decision-making and planning, millimeter-wave radar SLAM, and digital-twin-driven industrial diagnostics.

Topics: generative music, planning, radar SLAM, anomaly detection

Current projects

  • Code editing with LLMs Instruction tuning of large language models for code editing. — Li Kaixin (PhD)
  • Low-level vision Restoration and enhancement under adverse conditions. — Cao Xiao (PhD) with A/Prof Robby Tan
  • Software engineering on NLP Language-model methods for software engineering tasks. — Liu Chenyan (PhD) with Prof Dong Jinsong
  • Generative AI for music Generative models for musical content. — Gao Tianyu (PhD) with Prof Dong Jinsong
  • Decision making & planning Learning to plan and act under uncertainty. — Song Le (PhD) with A/Prof Bryan Low
  • Medical image computing Analysis and modeling of medical images. — Dai Yuhe (PhD) with Dr Huang Weimin
  • Radar-based SLAM Millimeter-wave radar based simultaneous localization and mapping. — Hao Zelin (PhD) with Prof Chen Xudong and Prof Guo Yongxin
  • Multilingual NLP Low-resource multilingual natural language processing. — Ai Xi (Postdoc) with A/Prof Min-Yen Kan