|January 31 · Issue #26 · View online |
As always, if you’d like to support deep learning weekly share this issue with friends and colleagues.
| Deep Learning Will Radically Change the Ways We Interact with Technology |
A nice introduction to deep learning, its applications and future developments.
| Global AI startup financing hit $5bn in 2016 |
AI is everywhere and the money is coming in. The article gives an overview of recent investments and tries to shed some light on the scene.
| Deep Learning for Computer Vision with Python |
Adrian Rosebrock is raising funds for his eBook, where he will deliver practical walkthroughs, hands-on tutorials and more.
| Learning Policies For Learning Policies — Meta Reinforcement Learning (RL²) in Tensorflow |
Get your feet wet with Meta Reinforcement Learning in this great introduction. Features rich visualisations and an implementation in TensorFlow.
| Deep Learning, Applied. |
Learn how to classify your food in one of 101 classes using a Keras/TensorFlow workflow.
| TensorFlow 1.0.0 - rc0 |
The first release candidate of TensorFlow 1.0.0 has been released and finally includes the Debugger, an initial release of the new domain-specific compiler and much more.
| Fully Convolutional Networks (FCNs) for Image Segmentation |
Daniil Pakhomov gives an introduction to his impressive library based on TensorFlow and TF-Slim. The library offers easy access to the Pascal VOC dataset, training pipelines and pretrained models.
| TensorLayer: Deep Learning and Reinforcement Learning Library for TensorFlow |
A powerful library developed at the Imperial College London, that offers a simpler interface to TensorFlow and simplifies building your models.
| Learning Light Transport the Reinforced Way |
Nvidia has come full circle and uses its graphics cards to learn new techniques for generating images on its graphics cards.
| Wasserstein GAN |
Facebook-Courant Institute researchers can get rid of mode collapse in GAN learning, as well as improve the stability of learning.
| Adversarial Variational Bayes: Unifying Variational Autoencoders and Generative Adversarial Networks |
This paper describes a technique for training Variational Autoencoders using an auxiliary discriminative network that promises higher quality results.