FastML

Machine learning made easy

Deep nets generating stuff

The last few weeks have been a time of neural nets generating stuff. By deep nets we mean recurrent and convolutional neural networks, while the stuff is text, music, images and even video.

Classifying text with bag-of-words: a tutorial

There is a Kaggle training competition where you attempt to classify text, specifically movie reviews. No other data - this is a perfect opportunity to do some experiments with text classification.

Kaggle has a tutorial for this contest which takes you through the popular bag-of-words approach, and a take at word2vec. The tutorial hardly represents best practices, most certainly to let the competitors improve on it easily. And that’s what we’ll do.

The emperor’s new clothes: distributed machine learning

We can think of two reasons for using distributed machine learning: because you have to (so much data), or because you want to (hoping it will be faster). Only the first reason is good.

Distributed computation generally is hard, because it adds an additional layer of complexity and communication overhead. The ideal case is scaling linearly with the number of nodes; that’s rarely the case. Emerging evidence shows that very often, one big machine, or even a laptop, outperforms a cluster.

What you wanted to know about AI, part II

In part one we attempted to show that fears of true AI have very little to do with present reality. That doesn’t stop people from believing: they say it might take many decades for machine intelligence to emerge.

How to dispute such claims? It is possible that real AI will appear. It’s also possible that a giant asteroid will hit the earth. Or a meteorite, or a comet. Maybe hostile aliens will land, there were a few movies about that too.

What you wanted to know about AI

Recently a number of famous people, including Bill Gates, Stephen Hawking and Elon Musk, warned everybody about the dangers of machine intelligence. You know, SkyNet. Terminators. The Matrix. HAL 9000. (Her would be OK probably, we haven’t seen that movie.) Better check that AI, then, maybe it’s the last moment to keep it at bay.

Torch vs Theano

Recently we took a look at Torch 7 and found its data ingestion facilities less than impressive. Torch’s biggest competitor seems to be Theano, a popular deep-learning framework for Python.

Loading data in Torch (is a mess)

Torch 7 is a GPU accelerated deep learning framework. It had been rather obscure until recent publicity caused by adoption by Facebook and DeepMind. This entirely anecdotal article describes our experiences trying to load some data in Torch. In short: it’s impossible, unless you’re dealing with images.

Interactive in-browser 3D visualization of datasets

In this post we’ll be looking at 3D visualization of various datasets using data-projector from Datacratic. The original demo didn’t impress us initially as much as it could, maybe because the data is synthetic - it shows a bunch of small spheres in rainbow colors. Real datasets look better.