Sander Dieleman

Actually Spotify does more than collaborative filtering. Here’s a superb blog post on using convolutional filters on the spectrograms to build content-based recommendations: https://benanne.github.io/2014/08/05/spotify-cnns.html

Highly recommended further reading: http://benanne.github.io/2014/08/05/spotify-cnns.html (Recommending music on Spotify with deep learning) uses CNNs trained on spectrograms + similarity data from collaborative-filtering to predict per-song vectors.

I don't have great links for this, but for something less technical, you might look at the blog posts from Kaggle competition winners. Here are a couple examples http://benanne.github.io/2015/03/17/plankton.html http://blog.kaggle.com/2016/04/13/diagnosing-heart-diseases-with-deep-… or check out what's available in a deep learning library like keras http://keras.io/preprocessing/image/ Sorry I don't have better academic references.

Last summer, Sander Dieleman was working on recommending music based on audio signals, while interning at Spotify. Very interesting work: https://benanne.github.io/2014/08/05/spotify-cnns.html

I noticed recently that the new discover weekly songs tend to sync perfectly with my walking speed. I wonder if I subconsciously leaked some bias into my training sample. BPM has to be one of the factors they are matching against. Here is a blog post that describes one of their machine learning models using neural networks. http://benanne.github.io/2014/08/05/spotify-cnns.html

Yes you are absolutely right, DSP knowledge is also a must. I didn't get that far, I started by identifying what makes people like a certain song and got it down to: Lyrical themes (and overall artist thematic) and melodic and rhythmic patterns. For instance by analyzing data from last.fm I found that a lot of people that listen to Black Metal also seem to have an interest in european/celtic/gaelic folk music, but not all only certain artists with specific lyrical themes and what seemed to me as specific melodic patterns. Anyway the thing that stumped me was identifying/creating the "features" for rhythmic & melodic patterns, eventually I realized that I was way out of my league and had to drop it. Later on I found that someone else was already working on this: http://benanne.github.io/2014/08/05/spotify-cnns.html

Here's a write-up about a Neural Net which was used to win a Kaggle image classification challenge, they did a lot of transformations on the input data to a) prevent overfitting b) provide invariance. Some other cool tricks mentioned there, too. https://benanne.github.io/2015/03/17/plankton.html

I am quite interested in this topic. There is some research on that field. Some links I collected in the past: http://thesis.flyingpudding.com/ http://dynampd.ubitux.fr/ http://erikbern.com/2015/09/22/presentations-about-spotify-music-recom… http://erikbern.com/2015/09/24/nearest-neighbor-methods-vector-models-… http://www.wired.com/underwire/2013/08/qq_netflix-algorithm/ http://www.slideshare.net/erikbern/collaborative-filtering-at-spotify-… http://benanne.github.io/2014/08/05/spotify-cnns.html https://news.ycombinator.com/item?id=8137264 http://www.playdar.org/ https://github.com/RJ/playdar-core http://news.ycombinator.com/item?id=3876724 https://github.com/bmcfee/librosa http://forever.fm/ http://blog.petersobot.com/introducing-forever-fm http://musicmachinery.com/2011/05/14/how-good-is-googles-instant-mix/ http://musicmachinery.com/2012/11/12/the-infinite-jukebox/ Btw., as I am developing my own music player, my main important feature/goal is a kind of automatic DJ which automatically selects music and I also especially had a discover mode in mind. However, that main feature did not evolve that much so far because there were so much other things to implement first. For now, it only supports access to files in your file system, and it looks on tags and artist and adds quite some randomness to it. https://github.com/albertz/music-player/blob/master/WhatIsAMusicPlayer…

This book is really coming together. It's been a while since I've put together a (100% not comprehensive) list of good places to start if you're looking to learn more and/or use deep learning in your projects. Open source Pylearn2 (used to win kaggle galaxies competition): http://deeplearning.net/software/pylearn2/ Theano (symbolic math library used by Pylearn2): http://deeplearning.net/software/theano/ Deep learning tutorials with theano (build your own neural networks): http://www.deeplearning.net/tutorial/ Demos Convnet JS: http://cs.stanford.edu/people/karpathy/convnetjs/ Sentiment Analysis: http://nlp.stanford.edu:8080/sentiment/rntnDemo.html 3d word cloud (webgl): http://wordcloud.ersatz1.com/ Commercial Ersatz (I'm a co-founder, it's a PaaS providing neural network software with cloud GPU servers): http://www.ersatzlabs.com Good Reading Deep learning of representations: looking forward http://arxiv.org/pdf/1305.0445v2.pdf Zero-Shot Learning Through Cross-Modal Transfer: http://arxiv.org/pdf/1301.3666v2.pdf <-- C'mon, that's pretty amazing... Solution for the Galaxy Zoo challenge: http://benanne.github.io/2014/04/05/galaxy-zoo.html Pylearn2 in practice: http://fastml.com/pylearn2-in-practice/