Continual Prototype Evolution: Learning Online from Non-Stationary Data Streams | Matthias De Lange

Paper Abstract:
The talk focuses on :
1) What is continual learning in machine learning?
2) How do neural networks learn from data streams that change over time?
3) Can we design a framework with minimal assumptions on the data?
4) We take a look at prototypes in representation learning, and use them for online learning from imbalanced and non-stationary data streams, much like in the real world!
Speaker Bio:
Matthias De Lange is currently a Ph.D. researcher @ KU Leuven (PSI)Interested in incremental and adaptive machine learning. He completed his M.Sc. “Engineering technology electronics and ICT, specialisation ICT” (Laureate 2018) @ KU Leuven. He is also General chair Continual Learning in Computer Vision workshop at CVPR 2021. Join us on June 25th: sites.google.com/view/clvisio....
Slides : mattdl.github.io/extra/2020_C...

[@Janamejaya.Channegowda]

Thank you for sharing the presentation, very useful, I was looking for resources related to continual learning, keep up the great work.