Stanford Seminar 2022 - DeepMind's Perceiver and Perceiver IO: New Data Family Architecture
Offered By: Stanford University via YouTube
Course Description
Overview
Explore a Stanford seminar on DeepMind's Perceiver and Perceiver IO architectures, presented by Senior Research Scientist Drew Jaegle. Delve into the development of flexible AI systems capable of processing diverse data types without domain-specific constraints. Learn how Perceivers utilize asymmetric attention mechanisms to handle high-dimensional inputs and outputs, outperforming domain-specific architectures across various tasks. Discover applications in image classification, natural language understanding, optical flow, multimodal processing, and multi-task learning. Gain insights into Jaegle's research on domain-general perceptual and reasoning systems, spanning architecture design, imitation learning, reinforcement learning, and self-supervised learning. Understand the potential impact of these innovations on the future of artificial intelligence and data processing.
Syllabus
CS25 I Stanford Seminar 2022 - DeepMind's Perceiver and Perceiver IO: new data family architecture
Taught by
Stanford Online
Tags
Related Courses
Stereo Vision, Dense Motion & TrackingUniversity at Buffalo via Coursera Computer Vision - Object Tracking with OpenCV and Python
Coursera Project Network via Coursera 3D Reconstruction - Multiple Viewpoints
Columbia University via Coursera Computer Vision for Visual Effects Lectures Spring 2014
Rensselaer Polytechnic Institute via YouTube Object Tracking and Motion Detection with Computer Vision
MathWorks via Coursera