In the future all of our automobiles, whether they are completely autonomous (ie, self-driving) or semi-autonomous (ie, partly controlled by us) will be connected.
This means one thing – that all of our cars will be, for intents and purposes, powered by data in many ways.
Which is why there is a thriving start-up scene looking at ways of making the most of the immense amounts of data that will be generated by the millions of cars on the world’s roads in the future.
One such company is Scale API, which trains artificial intelligence systems through the examination and categorization of visual data, who are passionate about the ways in which connected cars sharing data will mean a safer and more efficient road system for all of us in the future.
Scale API, in a nutshell, has developed automated systems that take data from self-driving cars see and apply labels to it, which then helps all autonomous cars make sense of the same situations re-occurring in the future.
They help driverless cars to learn, basically. And, of course, it’s no use a car learning in isolation. That’s why Scale’s customer base includes Cruise, Nuro, Lyft, Zoox, Nutonomy, Starsky Robotics, and Embark – all of whom send their data to Scale.
The nature of business means that connected car manufacturers are disinclined to share the data their cars are generating, which is the essence of the problem Scale has set out to fix.
Alexandr Wang, Scale’s 21-year-old founder and CEO, sums it up: “Right now each company is so in its own lane and secretive. In reality, these edge cases, these are things that should be probably be shared or standardized across the industry at some point.”
Self-driving vehicles produce a lot of data – from cameras, LiDAR, radar, and various other environments. That’s why Anantha Kancherla, who is in charge of developing autonomous software at Lyft told Wired:
“Scale is basically providing the ground truth for our perception systems. It’s a very, very critical piece for us to develop.”
Michael Wagner, CEO of Edge Case Research, added: “If you’re worried about your system missing edge cases, the ‘unknown unknowns,’ then the more examples you have, and the more conditions the car encounters, the more opportunities you have to train the system to do a better job.”
Sharing is caring. It’s also the way to a future in which self-driving cars are safer and way more reliable than they currently are.