NUS

NUS's AI research is carried by several core units:

• NUS AI Institute (NAII): the flagship institute launched in March 2024, organising foundational AI / applied AI / societal impact into three tracks
• School of Computing (SoC): traditional CS strength, with top teams across AI / NLP / CV / RL
• NUS Business School: AI in business applications, AI for Finance
• NUS Medicine: AI for Healthcare (in partnership with Synapxe and the public hospitals)
• NUS Law TRAIL: AI legal governance research

NUS also hosts AI Singapore — AISG's offices, compute, and faculty are all embedded in the NUS campus. This "national programme parasitically hosted in a university" model lets NUS enjoy both academic autonomy and national-level resources.

Internationally, NUS co-runs the IPP (Industrial PhD Programme) with Microsoft Research Asia, and also has research collaborations with Google DeepMind's Singapore lab.

NUS's AI research has a strong presence at all three levels:

• Foundational research: paper output at NeurIPS / ICML / ICLR / CVPR / ACL ranks in the top 3 in Asia, with several sub-areas (such as Trustworthy ML, Multi-modal Learning) in the global top 10
• Applied research: a large pipeline of horizontal projects with local enterprises, hospitals, and government agencies — NUS is the main "delivery vehicle" for applied AI research in Singapore
• National programme support: through AISG, NUS participates directly in national-level projects (SEA-LION training, AI Verify evaluation methodology, and so on)

Representative research directions:

• Trustworthy AI: teams led by Bryan Hooi, Reza Shokri and others are world-leading in AI privacy and adversarial robustness
• Multi-modal Foundation Models: the LV-NUS lab has consistent output on multi-modal LLMs
• AI for Science: collaborations with the biology, chemistry, and materials departments

But NUS's AI research has a long-running problem: talent drain to big tech and abroad — every year top PIs get poached by Google, OpenAI, and Meta. This is the structural challenge of a "small country with a high-cost environment" like Singapore.

NUS is the "load-bearing wall of foundational research" in Singapore's AI strategy.

Across the seven transmission levers:

• Lever 1 (Foundational Research): NUS is the single largest institution producing AI papers in Singapore
• Lever 2 (Talent): NUS Computing is the main pipeline for local AI engineers; about a third of AISG AIAP apprentices come from NUS
• Lever 3 (Industry Application): research is converted through NUS Enterprise and horizontal projects

Take: NUS is one of the few universities that can simultaneously produce world-class research and willingly take on national assignments — a balance that's rare in research universities. Embedding AISG inside NUS is the most critical institutional design in Singapore's AI strategy: it gives the national programme the research capacity of a university, and amplifies the university's research through national strategy.

But NUS AI Institute (founded 2024) is still in its integration phase. Whether it can truly become "Asia's AI research centre" needs another 2–3 years to judge. Key variables: ability to attract and retain top PIs, ability to produce SEA-LION-class flagship projects, and the ability to differentiate from NTU (NTU leans towards engineering applications, NUS towards foundational research).