ML4fun - From zero to Attention 4/5 - Alignment scores, context vector

Описание: "From zero to Attention", by Machine learning for fun (4 of 5)
Alignment scores, context vector

An introduction for starters to Neural Language Processing, which covers selected key concepts, up to the basics of Attention. No prerequisite knowledge required, we will cover everything from the ground up. Have FUN!

Complete set of videos
From zero to Attention 1/5 - Sequence models, encoders and decoders
   • ML4fun - From zero to Attention 1/5 - Sequ...  
From zero to Attention 2/5 - Neural Networks, RNNs
   • ML4fun - From zero to Attention 2/5 - Neur...  
From zero to Attention 3/5 - Embeddings, GRUs
   • ML4fun - From zero to Attention 3/5 - Embe...  
From zero to Attention 4/5 - Alignment scores, context vector
   • ML4fun - From zero to Attention 4/5 - Alig...  
From zero to Attention 5/5 - Performance, query, keys, values
   • ML4fun - From zero to Attention 5/5 - Perf...  




Reference paper: "Neural Machine Translation by Jointly Learning to Align and Translate", Dzmitry Bahdanau, Kyunghyun Cho, Yoshua Bengio - 2015 [https://arxiv.org/abs/1409.0473]

Cover: Photo by Power Lai on Unsplash