As an initiative supported by the Ministry of Business, Innovation and Employment (MBIE), TAIAO is committed to promoting and facilitating the adoption of artificial intelligence (AI) and data science in New Zealand's environmental sector.
In this newsletter, you’ll find more information about:
New Zealand International Olympiad team wins gold and silver
Web Search and Data Mining Conference 2026 (WSDM)
A brand new TAIAO talks episode
Learning opportunity - Agentic AI
What’s coming up
New Zealand International Olympiad team wins gold and silver
A huge congratulations to the New Zealand team (NZAIO) who brought home a gold and silver medal from the International AI Olympiad in Slovenia. The competition brought together top secondary-school talent from around the world to participate in complex AI challenges.
TAIAO was closely involved with Heitor Murilo Gomes (Chairperson of NZAIO) and Anton Lee (PHD student) supporting the New Zealand team throughout the competition. TAIAO was also able to support by sponsoring the team hoodies and t-shirts worn on the international stage.
A special acknowledgement to Vithya Yogarajan, the Executive Director of NZAIO, for supporting the students as one of the team leaders throughout the event.
To find out more about the team and the competition, head over to the NZAIO LinkedIn page (link to page).
Web Search and Data Mining Conference 2026 (WSDM)
On 18 February 2026, TAIAO PhD student Thomas Bailie took the international stage to present his research on HoGA: Higher-Order Graph Attention via Diversity-Aware k-Hop Sampling at the WSDM Conference in Boise, Idaho.
His research is a step toward representing influences from weather events happening far away and its impact on local weather, through a lightweight attention mechanism. It will provide a framework that can be used within TAIAO for directly representing teleconnections in models such as mesh-graphs, while also providing a benchmark for future work.
Thomas’ visit to Idaho was also an opportunity for him to connect with like-minded researchers and explore global work related to his field.
His work is supervised by TAIAO member, Dr Yun Sing Koh.
A new TAIAO talks episode
Our latest TAIAO talks episode is now live on the TAIAO website. It features Dr. Nick Lim as he introduces an AI powered flood prediction tool designed to help researchers forecast the impacts of future flood events with greater accuracy in Aotearoa.
Developed in July 2025, the tool is already making waves in the environmental data science space. In this episode, Dr. Lim breaks down how it works, why it matters, and the impact that this research has made.
Watch the full episode here.
New course on AI Education – Agentic AI
A new course is now available on the AI Education website offering a deep dive into the world of AI Agents. Just as CPUs transformed computation, Large Language Models are transforming intelligence. At the centre of this change are agents - AI systems that not only generate language but interact with the world on our behalf.
Lead by Prof. Albert Bifet, the course offers the opportunity to:
learn what makes agentic AI including how these systems not only generate answers but also sense, plan, and act
explore the opportunities agentic AI opens, the risks and ethical challenges it raises, and
see real-world examples that show how agents are beginning to transform work and industry.
The course is free and available on the AI Education website.
What’s coming up
International Conference on Learning Representations (ICLR)
PhD student Di Zhao is heading to Rio de Janeiro this April to present his accepted paper at the world‑leading ICLR conference.
Di’s research, supervised by Prof. Yun Sing Koh, focuses on building AI models that stay accurate in new or unseen environments (for example varying light conditions, different habitats, or changing weather). This challenge, known as “domain shift,” is a major barrier in environmental data analysis, where AI often struggles to handle Aotearoa’s constantly shifting environmental conditions.
Di’s research introduces a training approach that forces the model to “unlearn” misleading patterns it has picked up from a dataset or environment. He has been sponsored by TAIAO to attend the conference and showcase New Zealand’s growing strength in environmental AI on the global stage.
More details and photos to come in next month’s newsletter.
More episodes coming for TAIAO Talks
Weka.ai is TAIAO’s newest tool, and the focus for our next TAIAO Talks episode. It has been specifically designed to make learning about and using AI accessible for everyone.
In this episode, viewers will get a closer look at what weka.ai is, how it supports everyday learners, and the range of free educational resources it offers.
You can take a look at the weka.ai website here.
A new TAIAO podcast series
In our last newsletter, we let you know about the first series of our new TAIAO podcast on Quantum Machine Learning.
The TAIAO podcast will return next month with a second series on Quantum Machine Learning, hosted by World Bank Climate Challenge winner Léa Cassé. With five new episodes, this series will dive into quantum techniques for environmental data streams with an emphasis on rigour, reproducibility, and practical evaluation.
We’ll share the first episode in April, so stay tuned.
