Attention in Deep Learning and Physics Applications

US/Eastern
Description

Recorded Meeting Video: https://youtu.be/FMQFoFvurH0

    • 12:00 12:45
      Attention in Deep Learning and Physics Applications 45m

      Recently, so-called "attention mechanisms" have been added to deep learning architectures leading-to-state- of-the-art results in natural language processing using transformer architectures. In this seminar, we will present the fundamental principles behind attention mechanisms and show how attention-based architectures can be applied effectively to physics problems.

      Speaker: Pierre Baldi (University of California, Irvine)