NUS Libraries' AI Sense Maker Maps Four Centuries of Research in 30 Minutes
Source: AsiaOne
Finding the right academic paper in Singapore's university ecosystem has always been a bottleneck — not because the material isn't there, but because mapping four centuries of scholarship across disciplines takes time most researchers don't have. NUS Libraries is taking a direct shot at that problem with AI...
Finding the right academic paper in Singapore's university ecosystem has always been a bottleneck — not because the material isn't there, but because mapping four centuries of scholarship across disciplines takes time most researchers don't have. NUS Libraries is taking a direct shot at that problem with AI Sense Maker, a research-discovery platform that compresses two weeks of literature review into roughly 30 minutes by combining large language models with a curated knowledge graph of 150,000 digitised materials.
Built in-house over 12 months by NUS Libraries and NUS Information Technology, the platform accepts questions in conversational language and responds with a structured overview — a cited summary and an interactive concept map that surfaces adjacent ideas the user didn't know to search for. The underlying engine runs on OpenAI's large language model, but the differentiation comes from the taxonomy layer: NUS uses Wikidata as a starting point, then manually scopes and refines the relationships around its collections. That labour-intensive curation gives the model cleaner boundaries and keeps responses consistent across queries, a design choice that separates it from generic AI research assistants.
What makes this distinctive in Singapore's AI landscape is that it tackles a specifically local research pain point. General-purpose tools like ChatGPT or Gemini Notebook can summarise papers once you have them, but finding academically credible material — especially Singapore- and Southeast Asia-centric scholarship — still means iterating through search after search, chasing weak leads through ScholarBank@NUS and Digital Gems, the university's repository of rare and special collections. AI Sense Maker collapses that loop by pulling from both repositories, with references attached to every factual claim so researchers can trace answers back to sources. The oldest digitised books in its scope date to the 17th century.
The platform launches on 20 August 2026, first to NUS students and faculty, but the implications go beyond campus. If this approach proves effective — using managed taxonomies to give an LLM disciplinary boundaries — it offers a template for other Singapore institutions sitting on deep collections of regional knowledge. The National Library Board, A*STAR's research repositories, and the National Archives all hold material that would benefit from the same treatment. AI Sense Maker tests whether curated structure plus language model fluency can genuinely shorten the distance between a research question and a grounded starting point.
Why it matters for Singapore: The platform reinforces NUS's position as a regional AI research hub while addressing a concrete productivity gap in academia. It also demonstrates a model for how Singapore's institutions can apply AI to their unique knowledge assets — rare Southeast Asian texts, policy archives, scientific datasets — rather than relying solely on general-purpose tools built elsewhere. If the 20 August launch delivers on the 30-minute-from-question-to-bibliography promise, expect other Singapore universities and research bodies to follow suit.
