Behind the research: Guillaume Mougeot
Each week, we will introduce an interview with an early-career researcher working on the MAMBO project. Meet the scientists who bring innovation to the way we monitor biodiversity across Europe.
Can you introduce yourself and your research in a couple of sentences?
I'm Guillaume Mougeot, a postdoc and dreamer who has always wished to conduct research in every possible direction. My research interests today are mostly focused on deep learning and image analysis, as well as their application to biological and environmental questions. Thanks to the broad range of applications that image analysis enables, I have had the opportunity to work on topics such as animal re-identification, microscopy image analysis, and, more recently, insect detection and classification.
What knowledge gap is your PhD thesis focused on?
My research mainly focuses on training deep learning models to address still-unsolved image analysis problems in ecology. The two main ones I am currently working on are lepidoptera classification from camera trap images and invasive plant species detection from roadside images.
How has MAMBO directly shaped or enabled your thesis research? Were there datasets, fieldwork opportunities, or collaborations that changed your direction?
Insect species classification was one of the core goals of the MAMBO project, and this challenge gave me a great opportunity to collaborate with the PlantNet team, whose work on plant classification was closely related to mine. The diversity of MAMBO deployment sites also allowed me to evaluate our models across a wide taxonomic spectrum, covering insect species from Malta to Denmark. The PlantNet collaboration further opened new research directions in automated invasive plant species monitoring, including the development of a lightweight model for real-time monitoring during image recording.
Do your scientific results contribute to solving national or European problems, and can they be used to inform policies? If so, to which policies would they be relevant?
Insect classification models can be used by ecologists to extract relevant information about lepidoptera populations along three dimensions: time, space, and taxonomy. This could help build fine-grained knowledge of their conservation status, and assist in identifying both species in critical condition and the distribution of invasive species.
As MAMBO wraps up, how do you see its research to continue?
Through sister projects such as NextBON Horizon Europe project, the BEAGLE Horizon Europe project, or Biodiversa+ pilots, MAMBO's work could live on. Both the hardware systems and the software developed within MAMBO have much to offer the scientific community, serving as both practical tools and sources of new research opportunities. I also believe that the knowledge developed within MAMBO and shared through publications will lay the groundwork for new projects to emerge.
What's next for you professionally?
As a postdoc, I have greatly appreciated both the freedom to explore that MAMBO's broad goal of enhancing existing technologies has offered, and the rapid valorisation and real-world use of the tools I developed through the close partnerships within the project. I hope to continue this work in the future, ideally through a more permanent position.
Follow the early-career researchers campaign also on social media - LinkedIn, Bluesky and X.