🚀 It feels great to witness two years of hard work materializing into reality. Announcing Ego-Exo4D!
Video Summary: https://t.co/GJTDoJj5Bt
Website: https://t.co/yg3i7EV9nQ
Paper: https://t.co/8z4ZubWGzM
Video related to the paper: https://t.co/uJUEVeaSoe
@iiit_hyderabad
Today, we announce Ego-Exo4D, the largest and most diverse multi-view dataset, showing human experts around the world performing a core set of skilled activities, w/ unprecedented multi-modality, novel new video-language resources, and rich annotations
@eddybuild Hey @eddybuild, I have been following your work and got really interested in the potential that your datasets hold. What would be the best way to reach out to you out?
@MamataOfficial This comes as no surprise, given your leadership. The past 14 years have clearly reflected prolonged mismanagement, and it’s time for a change.
@gabriberton It at least holds true for I3D architecture with pseudo-labels on multiple video datasets, including ego-exo4d.. I have a paper on the same thing.. https://t.co/muuc2JAgXa
@gabriberton Another interesting paper - They show that the deep networks prioritize learning clean samples (low loss) over the noisy samples (high loss) in the initial epochs of training. https://t.co/9jnqIEbXVb
Ask ChatGPT
“based on what you know about me. draw a picture of what you think my current life looks like”
past your responses below.
thanks again @mreflow & @danshipper
Our academic partners use the Aria Research Kit to accelerate their research across CV and ML applications. @avijit9d and the CVIT lab at @iiit_hyderabad present their work on the Driver Intent Prediction project using Project Aria glasses.
Check it out: https://t.co/hNjm8Q3hjW