Zamba posts
Explore the latest in our mission to build a better world using data science and AI.
Content list
What's a takahē?
Although Zamba's models are trained with animals from Africa and Europe, they can be used with videos from other locations that show species the models have never seen. We demonstrate with a dataset from New Zealand.
Standing on the Threshold
Using probabilistic classifications from Zamba, we can automatically remove a large majority of blank videos while controlling the fraction of wildlife videos we lose. But how do we know where to draw the line?
Find all the pangolins
We can use Zamba's probablistic classifications to search for videos containing specific animals. Particularly for small animals, this strategy can be highly effective.
You can stop watching blank videos
Using Zamba's probablistic classifications, you can identify and remove blank videos -- saving viewing time, storage space, and data transfer costs -- while minimizing the loss of videos that contain animals.
Computer Vision for Wildlife Monitoring in a Changing Climate
Through the Patrick J. McGovern Foundation Accelerator, DrivenData and the Wild Chimpanzee Foundation are teaming up to create automated, accurate, and accessible species detection tools.
Join our newsletter or follow us for the latest on our social impact projects, data science competitions and open source work.
Protected by reCAPTCHA. The Google
Privacy Policy and
Terms of Service apply.
Work with us to build a better world
Learn more about how our team is bringing the transformative power of data science and AI to organizations tackling the world's biggest challenges.