News
🔜 [Aug 2024] Keynote & Tutorial at CCN
On Wednesday 7 August, Dhireesha Kudithipudi and I will give a keynote lecture and coding tutorial on continual learning at the annual conference on Cognitive Computational Neuroscience in Boston.
[May 2024] Michał presents PEC at ICLR
Our paper proposing Prediction Error-based Classification (PEC), a promising new approach for class-incremental learning, is published at ICLR 2024. Michał Zając, the first author of the paper, presented a poster at the conference.
[Apr 2024] Dagstuhl position paper accepted to TMLR
The position paper about why continual learning is important, which we wrote with participants to the Dagstuhl Seminar in March 2023, is now published in TMLR.
[Oct 2023] Continual AI Un-Conference
The first edition of the Continual AI Un-Conference is held on 19 October. Together with Julio Hurtado and Nikhil Churamani I organize the Call for Talks, an opportunity for researchers of all levels to present their innovative ideas, controversial views or unifying perspective to the community.
[Sep 2023] Award for Sergi’s Master thesis
Sergi Masip has been awarded the Best Master Project mention from the Master in Computer Vision (Barcelona). Sergi’s project, which I supervised together with Pau RodrĂguez, investigated continual learning of diffusion models. Congratulations Sergi!
[Jun 2023] 4th edition of the Continual Learning workshop at CVPR
This year I lead the organization of the Continual Learning in Computer Vision (CLVision) workshop at CVPR 2023. The workshop is held in Vancouver on Sunday 18 June.
[May 2023] Stability gap paper selected as ICLR spotlight
Our paper identifying the stability gap, led by Matthias De Lange, is published as notable paper (top 25%) at ICLR 2023.
[Mar 2023] Dagstuhl Seminar “Deep Continual Learning”
Together with Bing Liu, Vincenzo Lomonaco and Tinne Tuytelaars, I organized a Dagstuhl Seminar on Deep Continual Learning, which took place in Germany from Sunday 19 March to Friday 24 March.
[Dec 2022] “Three scenarios”-paper out in Nature Machine Intelligence
The article describing the three continual learning scenarios (task-, domain- and class-incremental learning) is published in the journal Nature Machine Intelligence. It has been a while since the preprint, and there are exciting new additions!
[Dec 2022] NeurIPS tutorial “Lifelong Learning Machines”
On Monday 5 December, together with Dhireesha Kudithipudi and Tyler Hayes, I presented an online tutorial on Lifelong Learning Machines at NeurIPS 2022.