
Helge Rhodin
@helgerhodin
Assistant Professor at UBC
ID: 1326284163243147265
10-11-2020 22:02:45
166 Tweet
1,1K Followers
79 Following

Day 1. Yesterday, our 1st keynote from @timnitGebru (@dair-community.social/bsky.social) covered fairness in computer vision, illustrating the danger of AI when applied to surveillance and segregation, and how to foster community-rooted computer vision research that has a social impact. #ACVSS


Day 2. Today, we were lucky to have Anurag Arnab lecturing about advanced architectures for vision. An exciting journey from MLPs to Transformer and their applications to recent MLLMs. #ACVSS



Day 3. Today, we had the privilege of learning from Abeba Birhane about the rise of AI and Computer Vision in surveillance. An exciting lecture with multiple workshops that challenged us to consider our own decisions regarding AI applications. #ACVSS


Day 4. Yet another great speaker. Big thanks to Karteek Alahari Karteek Alahari for lecturing on visual representation learning. So much knowledge was packed into a single lecture covering from incremental and multitask learning, all the way to continual learning. #ACVSS


Huge shoutout to the Ro'ya CV4Africa community where it all began during the Ro'ya DLI workshop! Thanks mennatullah siam and Helge Rhodin for leading also the mentorship session. The community support from Ro'Ya has been incredible. #ACVSS


Day 7. Camera Image Signal Processors (ISPs) are crucial to make image looks natural. We were lucky to welcome Mahmoud Afifi who gave a keynote on how to make learnable ISPs more reliable. #ACVSS


Day 8. It was about time to add the temporal dimension. Big thanks to Hazel Doughty (Hazel Doughty) for her exciting lecture on video understanding, covering methods from optical flow to transformers for video. #ACVSS


Day 8. Exciting keynote from Girmaw Abebe Tadesse (Girmaw Abebe Tadesse), principal scientist at Microsoft AI for Good, on their initiatives using AI to transform lives on the continent and the importance for AI makers to listen to people's needs. #ACVSS


Day 9. Big thanks to Sara Beery (Sara Beery) for her insightful lecture on reducing datasets bias and constructing deployable computer vision that, beyond benchmarks, also contribute to real-life conservation applications. Such an inspiring line of work. #ACVSS


Day 9. Nice keynote from Negar Rostamzadeh (Negar Rostamzadeh) on how to build towards responsible computer vision. She shared her views on the complex process of creating more transparent and accountable frameworks. #ACVSS





