On May 13th, IML senior researcher and CST group leader Dr. Thiago Gouvêa delivered a keynote at the Digital Green Talents Spring School 2025 in Steinbach, sharing insights into interactive machine learning for biodiversity conservation and his team’s achievements in the XPRIZE Rainforest competition.
The Spring School, part of the Digital Green Talents program initiated by Germany’s Federal Ministry of Education and Research (BMBF), brings together early-career scientists working on sustainable digital innovations. The 2025 edition focused on fostering cross-disciplinary dialogue and practical solutions for green tech transfer.
In his keynote, Thiago explored the role of human expertise in AI workflows, presenting a novel expert-in-the-loop machine learning framework designed to support species monitoring using sound recordings. Drawing on his experience in one of the top three teams in the XPRIZE Rainforest competition, he demonstrated how combining archived acoustic data, pre-trained models, and interactive tools enables experts to validate, refine, and even discover species in challenging field conditions. He also reflected on the limits of automation and the importance of structured, interpretable models that align with how domain experts reason and make decisions. The talk sparked thoughtful discussion, underscoring the importance of models that support, rather than replace, expert judgment—highlighting a key challenge for sustainable AI in complex environmental domains.
Digital GreenTalents Awardees with Dr. Rainer Müssner from the Federal Ministry of Research, Technology and Space
Dr. Thiago Gouvêa pitching a future where machines sift through the chaos—and experts get the interesting bits.
