Cookies
We use essential cookies to run the site. Analytics cookies are optional and help us improve World Wide. Learn more.
University of California Berkeley
Showing your local timezone
Schedule
Thursday, November 26, 2020
3:00 PM Europe/London
Recording provided by the organiser.
Domain
Host
Sheffield ML
Duration
70 minutes
While meta-learning algorithms are often viewed as algorithms that learn to learn, an alternative viewpoint frames meta-learning as inferring a hidden task variable from experience consisting of observations and rewards. From this perspective, learning to learn is learning to infer. This viewpoint can be useful in solving problems in meta-RL, which I’ll demonstrate through two examples: (1) enabling off-policy meta-learning, and (2) performing efficient meta-RL from image observations. I’ll also discuss how this perspective leads to an algorithm for few-shot image segmentation.
Kate Rakelly
University of California Berkeley
Contact & Resources
neuro
Digital Minds: Brain Development in the Age of Technology examines how our increasingly connected world shapes mental and cognitive health. From screen time and social media to virtual interactions, t
neuro
neuro
Alpha synuclein and Lrrk2 are key players in Parkinson's disease and related disorders, but their normal role has been confusing and controversial. Data from acute gene-editing based knockdown, follow