Posted in | Remote Monitoring

Using Artificial Intelligence to Study Wild Animals

Artificial intelligence is set to transform our world, from cancer detection to earthquake prediction to self-driving cars. The latest study sees researchers introduce the power of deep learning to a new domain—ecology.

A collaborative team of scientists from Harvard, Auburn University, the University of Wyoming, the University of Oxford and the University of Minnesota showed that the artificial intelligence method can be used to identify animal images caught by motion-sensing cameras.

Using over three million photographs from the citizen science project Snapshot Serengeti, scientists taught the system to automatically identify, count, and describe animals in their natural environments.

The results revealed that the system was able to automate the process for up to 99.3% of images as perfectly as human volunteers, and was recently published in the Proceedings of the National Academy of Sciences.

Snapshot Serengeti has installed several "camera traps," or motion-sensitive cameras in Tanzania to collect millions of images of animals in their natural habitats, such as cheetahs, leopards, lions, and elephants.

While the images can provide insight into numerous questions, for instance, how carnivore species co-exist or the predator-prey relationships, they are only beneficial once they have been adapted into data that can be processed.

For several years, the best technique for gathering such information was to request crowd sourced teams of human volunteers to label each image manually—anarduous and time-consuming process.

Not only does the artificial intelligence system tell you which of 48 different species of animal is present, it also tells you how many there are and what they are doing. It will tell you if they are eating, sleeping, if babies are present, etc."

Margaret Kosmala, Co-Author

A type of computational intelligence roughly inspired by how animal brains view and understand the world, deep learning depends on training neural networks using massive amounts of data. For that process to be effective, though, the training data must be correctly labeled.

First-author Mohammad Sadegh Norouzzadeh anticipates deep learning algorithms will continue to progress in the days to come and expects to see analogous systems applied to other ecological data sets.

We wanted to demonstrate the value of the technology to the wildlife ecology community, but we expect that as more people research how to improve deep learning for this application and publish their datasets, the sky's the limit. It is exciting to think of all the different ways this technology can help with our important scientific and conservation missions."

Mohammad Sadegh Norouzzadeh, First Author

This technology lets us accurately, unobtrusively, and inexpensively collect wildlife data, which could help catalyze the transformation of many fields of ecology, wildlife biology, zoology, conservation biology, and animal behavior into 'big data' sciences."

Jeff Clune, Senior Research Manager at Uber's Artificial Intelligence Labs

The paper was written by Clune, his PhD student Mohammad Sadegh Norouzzadeh, his former PhD student Anh Nguyen (currently at Auburn University), Margaret Kosmala (Harvard University), Ali Swanson (University of Oxford), and Meredith Palmer and Craig Packer (both from the University of Minnesota).

This study was financially supported by the National Science Foundation.

Tell Us What You Think

Do you have a review, update or anything you would like to add to this news story?

Leave your feedback