Artificial Intelligence

How do online and offline narratives—from political speeches to news coverage to social media talk— make and unmake a democracy?

Researchers from The University of Osaka compared psychological ratings of various words from humans and large language models (LLMs) along different dimensions in order to compare the ways in which they conceptualize words. Human and LLM ratings aligned closely for some attributes (such as concreteness) but diverged significantly for others (such as iconicity). This work reveals which linguistic features may be reliably estimated using LLMs.

Two-dimensional topological materials are widely studied because they can be efficient and cost-effective catalysts. Their unique electronic properties give rise to robust topological surface states. However, this assumption is often made under the premise that their surfaces remain clean and unchanged during reactions, which is rarely the case. Using quantum computing, a research team has recreated a 2D material’s true working surface under reaction conditions.

Sticky life beginnings. Precarious pitcher plant, Breaking the cobalt "cage", Toxic algae killer & “Pure-bred” stem cell medium. Read all in the latest Editor's Choice. Plus Asia Research News 2026 is out now and SciCom Coffee talk by Rachael Smith at Wellcome Sanger Institute.

• AI models can support doctors in making critical decisions and expand access to care in low-resource settings • However, gaps in infrastructure, skills and governance continue to limit safe and effective adoption • Researchers call for an international consortium to guide regulation and responsible use of AI in healthcare

A recent review article explores how software-based AI systems are changing the design process of solid electrolytes, enabling faster development of solid-state batteries.

What if that traffic associated with the daily commute could be put towards computing? It may sound like a stretch, but that is what researchers from Tohoku University have recently proposed, developing a new artificial intelligence framework that harnesses road traffic as a computing resource. The framework would lessen AI’s reliance on power-hungry hardware, making it more sustainable.

DGIST is creating AI that can physically collaborate with people, truly personalised healthcare, and quantum sensors which will perceive the world in entirely new ways.

The largest ever reported database of dielectric material properties could speed up development of electronics like smartphones and energy storage systems.

Understanding how tree-like structures that form in thin films could be the key to next-generation materials for beyond-5G communications technologies.

By identifying the ideal manufacturing conditions, machine learning reduces the need for expensive and time-consuming experimentation.

Lingnan University welcomed a delegation from the University Grants Committee (UGC) on its Tuen Mun campus on 8 January, and presented its latest developments as a research-oriented liberal arts university comprehensive in arts and sciences. The delegation toured Lingnan’s latest teaching facilities and viewed their research, speaking at length with senior management, academic staff, and students.

The Philippines is recruiting its first National AI Team to compete with students from over 60 countries in the annual IOAI.

Large language models accelerate construction of materials property databases.

An AI system that learns from experienced endoscopists and pathologists helps doctors in low-resource areas quickly check stomach health using standard endoscopy images.

Researchers at National Taiwan University developed an AI system that recognizes construction activities at both the individual and crew levels using ordinary site videos. The approach reveals how teamwork shapes productivity and provides a foundation for future human-robot collaboration on construction sites.

- A joint research team from DGIST and Sungkyunkwan University unlocked the secrets of semiconductor material growth using AI... Expecting new display and sensor technologies - Research results published in the Journal of the American Chemical Society (JACS), one of the world's most prestigious academic journals in chemistry

- DGIST-iProtein Therapeutics Joint Research Team innovatively enhances the binding strength and stability of IL-1Ra protein using supercomputing and structure-based design technologies. - The new variant 'E127Q' shows anti-inflammatory and neuroprotective effects that surpass existing treatments in cell and animal models.

Researchers at Osaka Metropolitan University developed models that classify X-ray images into specific body regions and simultaneously determine the imaging method and image orientation. Using these models, they successfully classified almost all data for use in deep-learning models.

Lingnan University is pleased to announce that Mr John Lee, Chief Executive of the Hong Kong SAR and Chancellor of Lingnan University, has re-appointed Mr Andrew Yao Cho-fai as Chairman of the Council and an ex-officio member of the Court for a further term of three years with effect from 1 January 2026.

A Filipino-led team of international researchers has developed a physics-based algorithm that could train AI to greatly improve weather forecasts.

Lingnan University participated in the two-day 5th Asia Exhibition of Innovations and Inventions Hong Kong (AEII), and delivered an outstanding performance. Among more than 140 innovative research projects presented by experts and inventors worldwide, all nine of Lingnan's participating projects received awards, securing two gold, four silver, and three bronze medals. This marks a remarkable leap from last year, when the University made its debut and achieved one gold and two silver medals.

Lingnan University congratulates elite student-athlete Yuen Wing-ki on winning a bronze medal for the Hong Kong Special Administrative Region (HKSAR) in the Women’s Table Tennis Doubles (Class TT5) at the 12th National Games for Persons with Disabilities and the 9th National Special Olympic Games (NGD). Her brilliant performance also sets a record for Lingnan’s outstanding sporting achievements.

Lingnan University’s Teaching and Learning Centre is proud to advise that its innovative IDEAL-Gen.AI (Instructional Design Enhanced Active Learning through Generative Artificial Intelligence) platform has garnered international recognition by winning the Bronze Award in the “Best Use of AI” category at the prestigious QS Reimagine Education Awards and Conference 2025.

A scientist has solved the conundrum of why robots struggle to pick tomatoes

GNSS Indonesia Summit & Expo 2025, the first international-scale GNSS forum in Indonesia, was successfully held from 19–23 November 2025 at Universitas Indonesia, Depok. This five-day event highlighted how global navigation satellite systems are becoming increasingly crucial across various sectors, including transportation, precision agriculture, disaster mitigation, environmental monitoring, and national digital transformation.

A research group led by Associate Professor Islam MD Parvez and Professor Kenji Hato of the Graduate School of Agriculture at Ehime University has developed "Hort-YOLO," a real-time monitoring system for horticultural crops. This system is an advanced object detection network focused on accurate and context-aware analysis of crops, and simultaneously achieves a balance between annotation speed and scalability of supervised learning, making it practical for deployment in real-world environments.

Prof Song Yongduan, Dean of the School of Data Science and Chair Professor of the Division of Industrial Data Science at Lingnan University, has been elected as a Foreign Member of the Chinese Academy of Engineering (CAE). The result was announced today (21 November), naming him one of the 24 newly elected foreign academicians for 2025. The honour recognises Prof Song’s outstanding contributions to control science and engineering at the international level.

When labelled scans are scarce and hospitals collect images in different ways, a new training recipe developed by SUTD researchers helps segmentation AI keep its bearings across domains without needing more annotations.

On November 17th, Lingnan University’s Teaching and Learning Centre (TLC) and The University of Sydney’s Educational Innovation (EI) Team co-hosted a pivotal seminar titled “The Transformation of Global Higher Education Learning Pedagogies: Harnessing the Innovation and Technology Revolution”. With an attendance of approximately 300 participants, both in person and online, the event highlighted the urgency to rethink and revolutionise educational practices to adapt to the AI era.