Lyft opens to the public autonomous driving data set from its Level 5 self-driving fleet

Lyft is offering to the public a set of autonomous driving data that it calls the “largest public data set of its kind,” containing over 55,000 3D frames of captured footage hand-labeled by human reviewers, data collected by seven cameras and as many as three lidars depending on the car used, plus a drivable surface map and HD spatial semantic data that corresponds to the captured info to provide context to researchers.

The data set is the work of Lyft Level 5, the division of the ride-hailing company that is responsible for its research and development of self-driving vehicle technology (hence the name, which is borrowed from SAE’s levels of autonomy for cars, with Level 5 being the highest). In a blog post detailing the move, Lyft notes that this part of the company has been working on its hardware and self-driving software for two years and wanted to make public some of the data it has collected in that time in order to “help level the playing field for all researchers interested in autonomous technology.”

Lyft Level 5 EVP of Autonomous Technology Luc Vincent notes in the post that the company is “committed to democratizing access to this tech,” and plans to continue to “release additional data” going forward. Lyft will also host a competition to inspire use of the data, with $25,000 in cash prizes and a presentation plus job interview opportunity at the NeurIPS conference this December.

That job interview opp is key to understanding why Lyft would want to make public this data it is collecting. Other autonomous driving companies have done similar things, including Aptiv for example, and there is indeed obviously some interest in sharing with other researchers in the field, as many of the people working at these companies come from a collaborative academic background. But it’s also a self-interested move, because it definitely helps these companies recruit talent in a hiring market that is competitive with few qualified individuals to go around.

Still, regardless of motivation, it’s definitely a net benefit to have data sets like this available to the public. Meanwhile, Lyft is currently working on its third generation of self-driving car, which includes a new sensor array with its own, proprietary camera with ultra-HDR capabilities.