How 3 software engineers built a record-breaking supernova identification system with deep learning

 

Pop into Dessa’s offices and you’ll soon find traces of the company’s fascination with outer space. A Lego replica of Saturn V, the rocket that made it to the moon, sits stationed on the reception area coffee table. Venture further, and you’ll discover that each of the meeting rooms are named after spaceports from around the world: Cape Canaveral, Kodiak, Guiana. For a few employees on Dessa’s software engineering team, the company-wide enthusiasm for space has evolved into a full-blown passion project called space2vec. 

Recently, the space2vec team has created a convolutional neural network that identifies images of supernovas with record-breaking speed and accuracy. Find the team’s full account of how they accomplished this over on our Medium page. 

 

How 3 software engineers built a record-breaking supernova identification system with deep learning

March 4, 2019