CORe: an Android app for Continual Learning & Object Recognition at the Edge!

Hello fellow Continual Learners! :smile:

I’m excited to present “CORe”: an android app for continual learning & object recognition at the edge we developed for commonly used android smartphones! :iphone::eye:

To the best of our knowledge, this is the first time a deep continual learning strategy is shown to be effective on such an highly constrained embedded device! :exploding_head:

We hope this will spark a new interest in continual learning at the edge, which has been always thought to be impractical for deep learning models before, but we believe it may constitute the backbone of pervasive and distributed AI computing in the future!

You can find more details on it below :point_down: Let me know what you think in the comments!

:round_pushpin: Medium Article: https://medium.com/continual-ai/latent-replay-for-real-time-continual-learing-at-the-edge-9a083c899856
:round_pushpin: Paper: https://arxiv.org/pdf/1912.01100.pdf
:round_pushpin: App repository: https://github.com/lrzpellegrini/CL-CORe-App

2 Likes

Great work. Where can I download this app?

Thanks @chenyanan! Here’s the app repo: https://github.com/lrzpellegrini/CL-CORe-App

1 Like

Great. Which method or algorithm is used to add a new class ?

1 Like

The method is called AR1* and it’s explained in details in this paper: https://arxiv.org/pdf/1912.01100.pdf