UH Hilo's data science lab becomes the Scientist-Centered AI Laboratory (SCAIL)
With a focused grant and a clear mission, the University of Hawaiʻi at Hilo's Data Science Laboratory is transforming into the new Scientist-Centered Artificial Intelligence Laboratory (SCAIL). The goal is simple and direct: build AI that helps scientists do better work-without replacing them.
"Data science is the best way to do science, if you're motivated by important real-world problems," says Travis Mandel, associate professor of computer science and coordinator for UH Hilo's data science program. Mandel founded UH Hilo's data science lab in 2017 and now leads SCAIL.
Mission: AI that keeps scientists in the loop
Mandel's expertise is in human-in-the-loop AI-an approach where researchers shape models through ongoing feedback and oversight. That keeps the science grounded and the models accountable. For background on the approach, see this primer from NIST on human-AI interaction best practices Human-in-the-loop.
As Mandel puts it, the lab is scientist-centered, not just science centered. The distinction matters: AI serves the work, and the researcher stays in control.
Cross-disciplinary work with measurable outcomes
The lab's collaboration model is straightforward: pair domain expertise with practical machine learning, then iterate with the scientists using it. Recent and ongoing projects include:
- Forest health and invasive species: With Professor of Geography and Environmental Science Ryan Perroy (Spatial Data Analysis and Visualization Lab), SCAIL applied computer vision to drone and helicopter imagery to identify Rapid ʻŌhiʻa Death (ROD) and other targets. Improved detection accuracy eased management and response planning.
- Coral mapping: With Associate Professor of Marine Science John Burns and the now-nonprofit MEGA Lab, the team supports large-scale coral mapping and monitoring workflows.
- Business analytics and new AI curriculum: With Associate Professor of Data Science and Business Sukhwa Hong and Assistant Professor of Quantitative Business Analysis Chenbo Shi, the lab supports applied AI in analytics and program development.
- Climate and ecosystems: Collaborations span soil moisture (Yinphan Tsang), cloud water interception (Tom Giambelluca and Han Tseng), and carbon sequestration with Christian Giardina, Director of the Institute of Pacific Islands Forestry (U.S. Department of Agriculture, Forest Service).
The through line: targeted models, clean data pipelines, rigorous validation, and interfaces scientists actually want to use.
New identity, same focus on practical value
A $5,000 grant from the Office of the Chancellor is funding the rebrand and a space upgrade on the second floor of Mookini Library. Expect ergonomic furniture, a standing desk, and a room designed for thinking, building, and collaborating-not just running experiments.
Small details matter. A lab people want to spend time in tends to produce better science.
Hands-on training for working scientists and future ones
Mandel's teaching style is direct and applied. Five undergraduates and a high school student currently work in the lab on projects that pair AI methods with real research questions. The work is challenging-by design-so students leave with problem-solving skills they can use immediately.
Opportunities aren't limited to computer science majors. The data science certificate requires about a year of coursework and gives students meaningful exposure to applied ML and data workflows. Many alumni land strong roles because they've already shipped real work.
Why this matters for research teams
- Scientist control: Human-in-the-loop cycles keep models aligned with field reality and evolving datasets.
- Broad applicability: From land use and ecology to business analytics, the same ML patterns (classification, segmentation, forecasting) can be adapted quickly.
- Operational value: Better detection, faster annotation, and clearer decision support reduce time from data to action.
If you're building similar capacity or upskilling your team, this resource maps the core skills and workflows: AI Learning Path for Research Scientists.
What's next
SCAIL is positioned as a campus hub for scientist-led AI. The lab's approach-tight collaboration, practical tooling, quick iteration-lowers the barrier for any research group ready to integrate ML into their workflow.
The signal is clear: serious AI work is happening here, and it's built around the people doing the science.
Your membership also unlocks:
