Professor Tanya Berger-Wolf, a world-renowned researcher at the intersection of artificial intelligence (AI) and ecology, is coming to U of T Engineering.
She will join the Edward S. Rogers Department of Electrical & Computer Engineering in the summer of 2027, where she will take up the Eddie Goldenberg Research Chair of Canada in AI for Nature.
Berger-Wolf describes herself as a computational ecologist, and her work leverages the power of data analysis to enhance our knowledge of ecology and biodiversity, and to support management and conservation decisions. Systems she has designed are used by organizations such as the International Union for Conservation of Nature (IUCN) to create management plans for endangered species.
“My PhD was in theoretical computer science, but my very first job as an undergraduate was in an ecology and evolutionary biology department, doing modelling and simulations,” she says.
“I’m also married to an ecologist, so I’m pretty fluent in the language of ecology.”
Among the early models and simulations that Berger-Wolf worked on was modeling the impact of environmental changes on various bird populations.
“The models we were building had more parameters than birds,” says Berger-Wolf.
“And for a lot of those parameters, we were essentially just guessing: what’s the birth rate? What’s the survival rate? You could get dozens of different answers depending on what assumptions you made, and I would often walk away thinking there just has to be a better way of doing this.”
That better way would emerge with the development of new sensing technology, including the small, yet powerful cameras that are now found in smartphones and many other devices. Berger-Wolf realized that the millions of nature photos taken every year — both for research projects using trail cameras and by nature enthusiasts all over the world — represent an incredibly rich data set, both for science and for conservation.
During her postdoctoral studies, Berger-Wolf worked with biologists such as Professors Simon Levin and Dan Rubenstein at Princeton University. One of her most cited papers focused on a study of Grévy’s zebras, of which only a few thousand wild individuals remain in Kenya and Ethiopia.
Berger-Wolf and her collaborators developed an AI system that could not only tell Grévy’s zebras apart from other zebra species but also identify individual animals from their markings.
Assembling data from hundreds of sightings, biologists and conservationists could extract crucial insights into behaviour: for example, which zebras are friends or how the mother-foal relationship evolves over time.
“Within a couple of months of publishing the paper, we had 77 requests from other researchers, saying ‘can you do my species next?’” says Berger-Wolf.
“That’s when we realized the power of AI for nature. Through photos, we can non-invasively — and at scale — do population size counts and even track individual behaviours. And on top of that, we reduce the use of radio collars or GPS tags, which are costly, cumbersome and can be stressful for these endangered species.”
The system eventually developed into Wildbook, an open-source platform that blends structured wildlife research with artificial intelligence, citizen science and computer vision. Now part of the non-profit Conservation X labs, the platform, now recognizing more than 250 species, continues to be used around the world to develop new insights to help fight extinction.
Currently, Berger-Wolf leads two major research centres. The first is the Imageomics Institute at The Ohio State University. Imageomics describes an emerging scientific field that brings together image analysis, machine learning and biological knowledge bases, analogous to other fields such as genomics or proteomics.
Using imageomics, researchers can extract information from images to gain insights into traits and relationships at individual, population and species scales — insights which in turn further improve the algorithms that run the tools.
For example, while many existing image identification models tend to focus on one specific class of organisms such as birds or trees, Berger-Wolf and her team have developed a powerful AI foundation model that covers the entire Tree of Life.
Known as BioCLIP, the most recent version was trained on over 210 million images of 1 million species, representing about half of all named species of plants, fungi, animals. Tests show BioCLIP outperforming many existing models, giving researchers around the world the ability to identify individual species across all these domains with high accuracy from a simple photograph.
“We also work on interpretable AI, which is capable of not just saying ‘this is that species of bird,’ but rather ‘this is that species of bird because it has a yellow belly, red wing, this shape of beak, and it wobbles when it walks,’” says Berger-Wolf.
At the same time, Berger-Wolf also leads the AI and Biodiversity Change (ABC) Global Center, which is jointly funded by the U.S. National Science Foundation and Canada’s Natural Sciences and Engineering Research Council. There, she and her international collaborators are developing multi-modal, multi-scale models to understand ecosystems and predict their response to global change.
Berger-Wolf says that the new research chair and her new role at U of T will take all of this work to the next level.
“As an AI researcher, I am very excited to be coming to one of the original homes of artificial intelligence,” she says.
“This was where Geoffrey Hinton did his critical work, and there are so many other researchers here who worked with him, and who I can collaborate with. I’m a very deeply interdisciplinary researcher, and what attracts me is the breadth of research going on here: no matter what challenge you’re trying to address, there is probably someone here working on it.”
“If we want to predict the state of ecosystems, and how they are going to respond to the pressures that we humans are putting on them, we need to get people from engineering, computing, biology, natural resources, geography and social sciences together. The survival of our own species depends on solving these challenges.”
