World Wide relies on analytics signals to operate securely and keep research services available. Accept to continue, or leave the site.
Review the Privacy Policy for details about analytics processing.
University of California Berkeley
Showing your local timezone
Schedule
Wednesday, November 25, 2020
4:00 PM Europe/London
Seminar location
No geocoded details are available for this content yet.
Recording provided by the organiser.
Format
Recorded Seminar
Recording
Available
Host
Sheffield ML
Seminar location
No geocoded details are available for this content yet.
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
neuro
The development of the iPS cell technology has revolutionized our ability to study development and diseases in defined in vitro cell culture systems. The talk will focus on Rett Syndrome and discuss t
neuro
Pluripotent cells, including embryonic stem (ES) and induced pluripotent stem (iPS) cells, are used to investigate the genetic and epigenetic underpinnings of human diseases such as Parkinson’s, Alzhe