HomeНаука и техникаRelated VideosMore From: Machine Learning at Microsoft

Sequence Modelling and NLP With Deep Learning (Keras)

57 ratings | 2381 views
Tim Scarfe takes you on a whirlwind tour of sequence modelling in deep learning using Keras! • Intro • Outline 2:03 • What is a neural network 2:38 • Concepts of deep learning 3:32 • What is a sequence? 8:34 • What is sequence processing? 9:28 • Tokenization 10:35 • word vectors vs word embeddings 12:06 • More about word embeddings 13:26 • Recurrent neural networks (RNNs) 15:26 • LSTMs 17:04 • GRUs vs LSTMs 18:31 • Bi-directional RNNs 19:28 • 1d CNNs and tour of convolutional filtering in MATLAB 20:22 • Stacking RNNs+CNNs 25:42 • Universal machine learning process 25:56 • Demo-1 hot encoding 29:17 • Demo-Defining RNNs in Keras 31:17 • Demo-IMDB in Keras 32:30 • Performance/scoring/eval of deep learning models 35:40 • Question on material and sigmoid activation 38:39 • Temperature forecasting problem (cover GRU, LSTM, regularisation, bidirectional, stacking) 41:55 • 1D CNNs 49.49 • Questions 52:00 Slides; https://github.com/ecsplendid/deep-learning-sequences-talk/blob/master/talk.pdf Make sure you buy yourself a copy of Francois Chollet's book https://www.manning.com/books/deep-learning-with-python
Html code for embedding videos on your blog
Text Comments (2)
Dinesh Mane (1 month ago)
Fantastic session Tim, and content is really good. You have explained return_sequence = True parameter with two examples which helped me to understand clearly. Thanks for sharing the video :)
awesome video sir. i will read the book.

Would you like to comment?

Join YouTube for a free account, or sign in if you are already a member.