About me

I am a PhD candidate at EPFL, Switzerland, supervised by Devis Tuia. My research focuses on applying recent machine learning methods to ecology and conservation. In particular, I study how deep learning can support large-scale biodiversity monitoring by improving species distribution models. As part of this work, I developed MaskSDM and co-developed CISO, two new deep learning frameworks for species distribution modeling. Recently, I proposed MIAM, a novel approach for handling modality imbalance and missing inputs in multimodal ecological applications. More broadly, while I remain critical of certain applications of AI, I am enthusiastic about the role it can play in monitoring the Earth and helping preserve its ecological diversity. In this regard, I am one of the organizers of the AI+Environment Summit, a full-day event designed to inspire, ignite, and innovate work in AI for the environment.

Before starting my PhD, I completed a Master’s degree in Data Science and a Bachelor’s degree in Communication Systems at EPFL. During my studies, I worked with several research groups, including the Data Science Lab (DLAB) of Robert West on analyzing food consumption at EPFL and the Machine Learning and Optimization Lab (MLO) of Martin Jaggi on text-to-image generation. I also worked with the INDY Lab of Patrick Thiran, where I co-built the first version of Climpact, which led to a publication. In addition, I interned at Sony in Stuttgart, working on neural architecture search. Alongside my research activities, I supervised several Master’s students and served as an assistant for multiple courses (see Teachings).

News

  • Feb 2026 — Our new masking strategy for multimodal ecological applications, called MIAM, has been accepted at ICLR 2026 🎉. Very proud of this work, which is a central contribution of my PhD. See you in Brazil 🪇🇧🇷
  • Feb 2026 — Our seasonal SDM approach with deep learning has been accepted for publication in Global Change Biology 🎊
  • Feb 2026 — I presented my work and career path as an AI engineer to high school students at both the Forum Formation in Sierre and the Forum Horizon at UNIL in Lausanne 🇨🇭
  • Jan 2026CISO has been published in Methods in Ecology and Evolution 🥳 I also presented this new approach for conditioning SDMs on incomplete observations at the TIBS conference in Aarhus 🇩🇰
  • Nov 2025 — Our MaskSDM paper has been published in Methods in Ecology and Evolution 🎉
  • Oct 2025 — I co-organized the AI+Environment Summit in Zürich 🌿

Selected publications

MIAM

MIAM: Modality Imbalance-Aware Masking for Multimodal Ecological Applications

Robin Zbinden*, Wesley Monteith-Finas*, Gencer Sumbul, Nina van Tiel, Chiara Vanalli, Devis Tuia
ICLR (2026)
🌐 Webpage  ·  📄 Paper  ·  💻 Code  ·  💽 Data

MaskSDM

MaskSDM with Shapley values to improve flexibility, robustness and explainability in species distribution modelling

Robin Zbinden, Nina van Tiel, Gencer Sumbul, Chiara Vanalli, Benjamin Kellenberger, Devis Tuia
Methods in Ecology and Evolution (2025)
🌐 Webpage  ·  📄 Paper  ·  💻 Code  ·  💽 Data

CISO

CISO: Species Distribution Modeling Conditioned on Incomplete Species Observations

Hager Radi Abdelwahed*, Mélisande Teng*, Robin Zbinden*, Laura Pollock, Hugo Larochelle, Devis Tuia, David Rolnick
Methods in Ecology and Evolution (2026)
🌐 Webpage  ·  📄 Paper  ·  💻 Code  ·  💽 Data  ·  🏞️ Poster

Spherical Harmonics and Sinusoidal Representation Networks

Geographic Location Encoding with Spherical Harmonics and Sinusoidal Representation Networks

Marc Rußwurm, Konstantin Klemmer, Esther Rolf, Robin Zbinden, Devis Tuia
ICLR (2024), ✨spotlight✨
🌐 Webpage  ·  📄 Paper  ·  💻 Code  ·  🏞️ Poster
🎬 Video

Pseudo-absences for species distribution modeling

On the selection and effectiveness of pseudo-absences for species distribution modeling with deep learning

Robin Zbinden, Nina van Tiel, Benjamin Kellenberger, Lloyd Hughes, Devis Tuia
Ecological Informatics (2024)
🌐 Webpage  ·  📄 Paper  ·  💻 Code

See all publications

Contact me

Feel free to reach out to me at robin.zbinden[at]epfl.ch, or come say hi at conferences. I’m always happy to discuss research ideas, potential collaborations, or anything related to machine learning and ecology — and I also enjoy chatting about sports, literature, films, music, or politics.