CMU Neural Nets for NLP 2021 (7): Attention

preview_player
Показать описание
This lecture (by Graham Neubig) for CMU CS 11-747, Neural Networks for NLP (Spring 2021) covers:

* Attention
* A Case Study: "Attention is All You Need"
* Improvements to Attention
* Specialized Attention Varieties
* What do We Attend To?

Рекомендации по теме
Комментарии
Автор

You know there are fast transform based neural networks where you swap what is adjusted. You use fixed dot products (statistical summary measures and filters) and adjustable (parametric) activation functions that modulate the responses of the dot products. AI462 neural networks.

hoaxuan