3D Faces Generated From 2D Photos, Machines Learning to Hand-Write & More

Transmission #6

3D face models generated from a single 2D photo! Sign up for updates

This week’s newsletter includes 3D faces generated from 2D photos using deep learning, a deep learning model learning to hand-write, an open source deep learning curriculum, driverless vehicles regulation in Michigan and more. Enjoy!

Each week I (@olivercameron) will be sharing the very latest news in deep learning and self-driving cars. To get priority access to these newsletters, please join the mailing list at transmission.ai!


Photorealistic Facial Texture Inference Using Deep Neural Networks

Researchers at Pinscreen have published a paper on face digitization. They present a data-driven inference method that can synthesize a photorealistic texture map of a complete 3D face model given only a partial 2D view of a person in the wild. Crazy! Read the paper…

Udacity Self-Driving Car Event

Udacity welcomed over 400 self-driving car enthusiasts to a private event in Mountain View, and videos from the talks have been uploaded to YouTube. There’s an update on the Udacity open source self-driving car project, an awesome Q&A between Sebastian Thrun and Axel Gern, and a fantastic talk from George Hotz of Comma on their new project Neo. Enjoy!

Deep Learning Reinvents the Hearing Aid

A great read from a group of researchers at Ohio State University who are aiming to achieve a 10x improvement on the standard hearing aid using deep learning. Such efforts could have transformational impacts for those who are hearing-impaired (~15% of adults). Be sure to listen to the audio! Read more…

Four Experiments in Handwriting with a Neural Network

Four fun deep learning experiments focussed on handwriting, created by members of the Google Brain team. Set aside 15 minutes to play with the interactive visualizations, including a deep learning model being asked to handwrite text. Read more…

Open Source Deep Learning Curriculum

This open-source deep learning curriculum is meant to be a starting point for everyone interested in seriously studying the field. The author strove to put a list of resources that form a logical progression from fundamental to advanced. Enjoy!

New Michigan Law Prepares for Autonomous Future

This is a big deal! The law specifically allows vehicles that have no human controls — no steering wheel, no pedals — to be tested in Michigan. It makes clear that both automotive companies and tech companies are able to operate self-driving ridesharing vehicles. And it allows the sale of autonomous vehicles to the public once the technology has been tested and certified. Read more…

The Major Advancements in Deep Learning in 2016

A worthwhile read on the techniques and tools that have come to the forefront in deep learning in 2016. The section on Generative Adversarial Networks is great! Read more…

Recurrent Dreams And Filling In

A fun write-up on playing with generative models that focus on infusing motion into static pictures. The results, with the relatively small amount of data, are promising! Read more…


That’s it for this week, thanks for reading! If you have any thoughts or questions, I’d love to hear from you in Tweet-form. You can follow and message me at @olivercameron.

Don’t forget, to get priority access to these newsletters, please join the mailing list by visiting transmission.ai!