Machine Learning Resources

I am currently working on audio/speech processing, TTS etc., so…

Deep Learning in Speech resources:

• Understanding sound/audio:
  • @radekosmulski - A great starting point, also usable in Google Colab.
  • Very good intro to audio.
  • Progress tracker in NLP by @seb_ruder
  • Everything you should know about sound

• Generation/Synthesis of new sounds based on training set (copied from):

  • Jake Fiala: “Deep Learning and Sound” http://fiala.uk/notes/deep-learning-and-sound-01-intro
  • GRUV: https://github.com/MattVitelli/GRUV. Btw, found that LSTM worked better than GRU.
  • John Glover: http://www.johnglover.net/blog/generating-sound-with-rnns.html Glover used LSTM fed by phase vocoder (really just STFT).
  • Google Magenta for MIDI: https://magenta.tensorflow.org/welcome-to-magenta
  • Google WaveNet for Audio… https://deepmind.com/blog/wavenet-generative-model-raw-audio/
    • WaveNet is slow. “Fast Wavenet”: https://github.com/tomlepaine/fast-wavenet
    • WaveNet in Keras: https://github.com/basveeling/wavenet
• Neural voice cloning with few samples implementations (to be added):
  • Transfer Learning from Speaker Verification to Multispeaker Text-To-Speech Synthesis - paper
    • Github implementation
  • Neural Voice Cloning with a Few Samples - paper
    • Github implementation
  • DeepVoice3 - paper
  • Wavenet - blog
• Text-to-speech implementations (to be added):
  • Tacotron2 - paper
  • DCTTS - paper

Deep Learning cheatsheet by @rusty1s.

Resources by @thom_wolf.

My self-educational approach is usually to get a few rather exhaustive books and read them for cover to cover.

• Here is my reading list to join the NLP/AI/ML field.
  • The “Deep Learning” Book by Ian Goodfellow, Yoshua Bengio and Aaron Courville is a good ressource to get a quick overview of the current tools.
  • “Artificial Intelligence: A Modern Approach” by Stuart Russell and Peter Norvig is a great ressource for all pre-neural-network tools and methods.
  • “Machine Learning: A Probabilistic Perspective” by Kevin P. Murphy is a great ressource to go deeper in the probabilistic approach and get a good exposure to Bayesian tools.
  • “Information Theory, Inference and Learning Algorithms” by David MacKay is a little gem that explain propabilities and Information theory so clearly it’s almost unbelievable.
  • “The Book of Why: The New Science of Cause and Effect” by Pearl, Judea is a good introduction to Causality (more accessible than the big “Causality: Models, Reasoning and Inference”)
  • “Reinforcement Learning: An Introduction” by Richard S. Sutton and Andrew G. Barto is a great ressource to get an introductory exposure to Reinforcement Learning
• Natural Language Processing: three great ressources I’ve read with interest:
  • Kyunghyun Cho’s lecture notes on “Natural Language Processing with Representation Learning” are great.
  • Yoav Goldberg’s book on “Neural Network Methods in Natural Language Processing” is nice too (see also an older free version here).
  • Jacob Eisenstein’s textbook on “Natural Language Processing” is also a very exhaustive read.
• It’s also good to complement this with a few online courses depending on what field you feel you should be diving deeper into. I took the following classes:
  • Computational Probability and Inference (6.008.1x) from edx.
  • Probabilistic Graphical Models Specialization from coursera.