Research Themes – NUS AI Lab

RESEARCH

We work on almost all areas of AI foundations including modeling and representation, reasoning and planning, machine learning and data science, computer vision and natural language processing as well as learning theory.

Embodied AI

A general intelligent agent needs to act effectively in the real world. Embodiment is also arguably necessary for achieving general intelligence. Our research into embodiment include work on mobile robots, autonomous cars, drones, as well as on manipulation in real physical and virtual simulated environments.

Interactive AI

In our vision, AI should work for people, and alongside people. AI should be able to understand people in order to help us achieve our goals. When humans and AI are able to work well together, the effectiveness often multiplies, outperforming what is achievable otherwise.

Trustworthy AI

AI systems are being deployed at an accelerating rate. With deployment, it is our responsibility to ensure that human lives are not adversely affected. Issues such as fairness, transparency, accountability, explainability and robustness are part of our main focus.

Projects

Intelligent Systems in Balance Sheet Forecasting

AI That Understand Human Emotions and Intentions

Occupational Impact of AI

Smart Cognitive Impairment Screening Tool

Integrating Deep Learning, Statistical Models and Structured Representations

Transfer Learning in Dynamic Environments

Rationalizing Deep Learning Model Decisions

Neuralizing Algorithms

Bayesian Optimization Lower Bounds

Tactile Learning & Perception

Artificial Theories of Mind & Body

Decentralized Federated Learning

Self-Supervised Hand Pose Estimation

Clinical Data Analytics

Implications of Real-time Competition in Online Markets

Retina Images for Diabetic Retinopathy and Glaucoma