Sharing some of the latest research, announcements, and resources on deep learning.
By Isaac Madan
Continuing our series of deep learning updates, we pulled together some of the awesome resources that have emerged since our last post. In case you missed it, here are our past updates: March part 1, November, September part 2 & October part 1, September part 1, August part 2, August part 1, July part 2, July part 1, June, and the original set of 20+ resources we outlined in April. As always, this list is not comprehensive, so let us know if there’s something we should add, or if you’re interested in discussing this area further.
Research & Announcements
Detecting Cancer Metastases on Gigapixel Pathology Images by Google Brain. Google demonstrates a CNN deep learning architecture for identifying tumors in pathology images, at a level better than a human doctor.
Introducing Keras 2 by Francois Chollet. The author of Keras announces its rapid ascension from one to over 100K users in two years, as well as their official 2.0 release with a new API and tighter TensorFlow integration.
AudioSet by Google Research. A large-scale dataset of manually annotated audio events: 632 audio event classes and a collection of 2,084,320 human-labeled 10-second sound clips drawn from YouTube videos. Useful for audio event detection applications.
Deep Voice: Real-Time Neural Text-to-Speech for Production by Baidu Research. A production-quality text-to-speech system constructed entirely from deep neural networks. Deep Voice lays the groundwork for truly end-to-end neural speech synthesis. Baidu also launches SwiftScribe for automated audio transcription here.
Overcoming catastrophic forgetting in neural networks by DeepMind. Researchers enable networks to learn tasks in a sequential fashion, as a human would, rather than forgetting prior tasks. Traditionally, neural networks are not capable of this — they experience “catastrophic forgetting.”
Fundamentals of Machine Learning Workshop at Stanford. Full day workshop on Friday March 31st, 2017 from 9:00am to 4:45pm on the basics behind modern machine learning algorithms, spanning supervised learning, unsupervised learning, ensemble methods, deep learning, and more. The workshop is open to the public.
Bayesian Deep Learning Workshop NIPS 2016. Excellent videos from this workshop at NIPS in December, spanning an overview of Bayesian Neural Network history, applications, open research questions, and much more.
TensorFlow Image Recognition on a Raspberry Pi by Matthew Rubashkin. Part 4 of broader series started in September 2016, Introduction to Trainspotting — a walkthrough by Silicon Valley Data Science on using deep learning to build a train detector. This post demonstrates using TensorFlow on a Raspberry Pi to do image recognition.
Deep Learning Resources Matrix by Bob Kot. A cheatsheet that describes basic info on deep learning frameworks like TensorFlow and Theano.
Deep Learning for Natural Language Processing course at Oxford. Advanced course on natural language processing, focusing on recent advances in analyzing and generating speech and text using recurrent neural networks. Many lecturers from DeepMind. Course materials are available on GitHub here.
The Black Magic of Deep Learning — Tips and Tricks for the practitionerby Nikolas Markou. A helpful cheatsheet on valuable experiential advice & insights gleaned from working with DNNs in practice. Great things to consider as you leverage deep learning.
Applying Machine Learning To March Madness by Adit Deshpande. Interesting and timely post on using machine learning to predict game outcomes for the upcoming NCAA tournament.
A Guide to Deep Learning by YerevaNN. A well-organized and well-designed guide to deep learning, outlining and ranking important resources and papers to learn the subject effectively.