Sparse Word Embeddings
In this post, I’ll give a brief introduction to word embeddings, how they’re trained, why sparsity is important and how we can induce it.
Recent posts
Machine Learning: Starting from Zero
Just over a year and a half ago I was sitting at a desk trading stocks, knowing close to nothing about Machine Learning and having last coded in a bit of Mat...
Sentiment Analysis with Transformers
The Transformer is the latest advance in Deep Learning architectures that has driven most state-of-the-art progress in NLP since it was first presented in ‘A...
Visualising CNN feature-maps and layer activations
Convolutional Neural Networks are the most successful deep learning architecture for Computer Vision tasks, particularly image classification. They comprise ...