Oxford University Machine Learning Course | Hacker News Reader

Oxford University Machine Learning Course

cs.ox.ac.uk

185 pointsjcr18 comments

The best part about this ML course is that all assignments are in Torch (a deep learning framework in Lua) for which Andrej Karpathy has good things to say on his blog[0]

> "Brief digression. The code is written in Torch 7, which has recently become my favorite deep learning framework. I've only started working with Torch/LUA over the last few months and it hasn't been easy (I spent a good amount of time digging through the raw Torch code on Github and asking questions on their gitter to get things done), but once you get a hang of things it offers a lot of flexibility and speed. I've also worked with Caffe and Theano in the past and I believe Torch, while not perfect, gets its levels of abstraction and philosophy right better than others."

[0] - http://karpathy.github.io/2015/05/21/rnn-effectiveness/

Just to promote a heterodox(ford) opinion: while Torch is great, learning Theano might be a better use of a budding data scientist's time because python is used much more widely in the industry.

I think learning a new language like Lua along with Torch is probably useful if someone is doing cutting edge neural network research.

And to promote an even heterodoxer opinion, learning Deeplearning4j might be an even better use because Java is the language of corporate IT, the language of Hadoop (and therefore big data), and because Spark is based on the JVM. If budding data scientists want their neural networks to scale, they should think about distributed systems from day 1. http://deeplearning4j.org

Full disclosure: I helped create DL4J, and it is a younger framework than both Theano and Torch.

Definitely, but I think Oxford University is more interested in teaching the underlying principles than tracking what's popular in industry.

The approach generally taken by Oxford and similar is that a scripting language like Python or Lua is simple enough (or at least the bits of it needed for the course are) that anyone taking the course can pick it up with no specific learning necessary.

I've found using nolearn and lasagne on top of Theano made Theano much easier to use while still being in Python with access to familiar graphics routines.

lasagne gives you ways of constructing neural network layers (implemented as Theano functions).

nolearn sits on top of lasagne and gives a Scikit learn style interface that makes it trivial to set up a standard deep network to predict values from given input data.

Using nolearn was a very similar experience for me to using the Torch7 framework.

Here's a radical opinion: obsessing over frameworks is counterproductive; you most certainly should not be evaluating a class on a subject that is so math/concept-intensive by the class's choice of implementation coding framework. Especially since many of these frameworks are more alike than different, in the grand scheme of things.

What's the barrier to entry like for this?

For example, I've never been particularly great at math.

"Machine Learning is a mathematical discipline, and students will benefit from a good background in probability, linear algebra and calculus. Programming experience is essential."

Check Prerequisites: https://www.cs.ox.ac.uk/teaching/courses/2014-2015/ml/index....

Khan academy is a great resource for remediating your math skills. I went through to cover all the last-third-of-the-textbook concepts that I never learned in school. To get good at something you just have to go do it, math included.

Having looked at the problem sheets, probability (think distributions), linear algebra and calculus are a must - Khan Academy is a great resource but nothing really tests you like university level homework! There was a brilliant Prob/Stats course on iTunesU that got me through 3rd year - worth a look.

Can you share a link/name please?

Joe Blitzstein's Probability course at Harvard (Stat110) is very well regarded. http://projects.iq.harvard.edu/stat110/home

if you've had college level linear algebra, probability, calc before, there's lots of good review/reference materials

https://www.reddit.com/r/MachineLearning/comments/1jeawf/mac...

https://www.metacademy.org/roadmaps/cjrd/level-up-your-ml

I think the Prob books by Sheldon Ross are good tho there's negative Amazon reviews, and the three LA texts by Axler, Strang and Insel/Friedberg/spence are worth buying (older editions for < $25 shd be good enough

Professors teaching such courses at the graduate level primarily intend them to benefit students doing active research in the field. There's an extraordinary emphasis on mathematical derivations to show how one idea leads to another. This is intended to both (1) provide insight into why methods work—though less rigorous and more intuitive approaches frequently exist and are discovered by the less robotic students, and (2) give students practice in the mathematical gymnastics needed to publish in the field. The benefit to a practitioner, let alone an interested outsider, will likely be small.

An analogy might be to consider whether a course on type inference and Hindley-Milner offered by a computer science department would benefit someone interested in learning Haskell.

I don't get it. I can't recall a single piece of learning content from Oxford that didn't have the audio quality of a tin can phone. I sounds like an over-compression issue, but it is just made even worse by horrible audio in what seems like tiny little boxes that lectures are held in. Someone please point Oxford towards a course on audio recording.

This. It's a shame that such promising lecture vids are close to inaudible.

I took Nando's ML courses at UBC 2 years ago. He's great at explaining complex concepts in digestible chunks. He's able to show how ML theories are modeled after natural processes well too (such as how speech recognition and image processing work using deep learning and neural nets).

Reply on news.ycombinator.com