In past posts, I’ve described the geometry of artificial neural networks by thinking of the output from each neuron in the network as defining a probability density function on the space of input vectors. This is useful for understanding how a single neuron combines the outputs of other neurons to form a more complex shape. However, it’s often useful to think about how multiple neurons behave at the same time, particularly for networks that are defined by successive layers of neurons. For such networks – which turn out to be the vast majority of networks in practice – it’s useful to think about how the set of outputs from each layer determine the set of outputs of the next layer. In this post, I want to discuss how we can think about this in terms of linear transformations (via matrices) and how this idea leads to a tool called word embeddings, the most popular of which is probably word2vec.

Zoubin Ghahramani has pointed out the odd nature of the term “Artificial Intelligence”. As an example he says it would be non-sensical to talk about artificial flying. Things either fly or they don’t.

I have been thinking about doing a “Becoming a Data Scientist” podcast for a long time, at least since April. The podcast would include interviews focused on how people working in various data-science-related jobs got to where they are today (how did they “become a data scientist”?). I’m getting closer to taking the dive and getting it started.

In my last blog I showed how to use convolutional neural networks to build a model that removed stains from an image. While convolutional neural networks seem to be well suited for image processing, in this competition I found that deep neural networks performed better. In this blog I show how to build these models.

When we hear about Convolutional Neural Network (CNNs), we typically think of Computer Vision. CNNs were responsible for major breakthroughs in Image Classification and are the core of most Computer Vision systems today, from Facebook’s automated photo tagging to self-driving cars.

### The Deep Learning Gold Rush of 2015

In the previous example, we have seen how to set up a simple Torch application to recognize digits from the MNIST set.

** Mon 02 November 2015

** Mon 02 November 2015

I’ve written before about how much I enjoyed Andrew Ng’s Coursera Machine Learning course. However, I also mentioned that I thought the course to be lacking a bit in the area of recommender systems. After learning basic models for regression and classification, recommmender systems likely complete the triumvirate of machine learning pillars for data science.