Self-Driving Shuttles, Model 3 Spy Shots, Photonic Neural Networks, Lip Reading with Deep Learning & More
Transmission #3
This week’s newsletter includes a full self-driving demonstration from Tesla, Model 3 photos you haven’t seen, a self-driving shuttle on campus, an open source by-wire control kit, photonic neural networks, lip reading in the wild, a discussion on what makes machine learning hard and a new deep learning framework for Apple devices.
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!
Video of the Week 📹 A Self-Driving Tesla Model X
The Photos of the Tesla Model 3 You Haven’t Seen
Speaking of Tesla, the Model 3 might just be the first fullyself-driving car available to the mass market. Many of the autonomy enthusiasts I know are one of the 400,000 people to have Model 3 on pre-order, but have yet to actually seethe car in anything but press shots. It turns out there’s tons of Model 3 spy shots out there, so I curated a bunch. View the photos here…
Self-Driving Shuttle Startup Auro Launches
The key to building a successful self-driving vehicle startup is to ship. The more you ship, the more data you can collect. Last week, I was excited to see my friends at Auro Robotics launch their self-driving shuttle at Santa Clara University. Read more about their launch. Oh, and they’re hiring!
The OSCC Project
The Open Source Car Control Project enables engineers to build their own self-driving development vehicle using existing by-wire technologies on the 2014-or-later Kia Soul. OSCC can be integrated into a new or used vehicle for less than $1,000. I’m super excited to see what this enables in the industry. Read more…
World’s First Photonic Neural Network Unveiled
Early results, but exciting! “The results show just how fast photonic neural nets can be. “The effective hardware acceleration factor of the photonic neural network is estimated to be 1,960 × in this task,” say Tait and co. That’s a speed up of three orders of magnitude.” Read more…
Lip Reading Sentences in the Wild
University of Oxford researchers continue to tackle the problem of reading lips using deep learning, and have published two potentially novel contributions: first, a pipeline for fully automated large-scale data collection from TV broadcasts; second, CNN architectures that are able to effectively learn and recognize hundreds of words from this large-scale dataset. Read the paper…
Why is Machine Learning ‘Hard’?
An excellent post by @zaydenam, discussing just how different machine learning is to traditional software engineering: “What is unique about machine learning is that it is ‘exponentially’ harder to figure out what is wrong when things don’t work as expected” Read more…
An Open Source Deep Learning Framework for iOS, OS X and tvOS
The goal of DeepLearningKit is to support using pre-trained Deep Learning models on all Apple’s devices that have GPUs. It is developed in Swift and Metal — to efficiently use on-device GPU to ensure low-latency Deep Learning calculations. DeepLearningKit sounds very useful! 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.
