Waymo is calling all intrigued minds to partake in an Open Dataset challenge, and it could result in some cold, hard cash.
On Thursday, Alphabet’s self-driving car subsidiary added more to its Open Dataset and invited anyone to tackle one of five challenges related to the data. The Open Dataset is a treasure trove of data that Waymo’s self-driving car prototypes collect as they log millions of miles on the road. Last year, the company shared the data and let anyone access it for the greater good of scientific progress.
With the latest data dump, it’s doing things a little different and gamifying things a tad. Researchers can participate in one of the five challenges: 2D detection, 2D tracking, 3D detection, 3D tracking and Domain Adaptation.
The 2D challenges require researchers to 2D boxes for objects in a scene, when given a set of camera images or a specific sequence in the case of 2D tracking. The 3D challenges are similar but include lidar range images to analyze and 3D upright boxes for the images or specific sequence. Domain Adaptation includes numerous segments from Waymo vehicles operating in rainy conditions.
Waymo will let teams include up to 10 people and they may submit their work up to three times. First place will walk away with $15,000 — not too shabby. Second place will receive $5,000 and third place will nab $2,000.
The challenge starts today and runs through May 31, and Waymo will keep a scoreboard of top performing teams. Waymo workers will be included in the scores, but they won’t receive any cash prizes, so don’t be discouraged.