BROADCASTS.com

One of the hardest problems in the automated driving stack involves predicting how multiple pedestrian targets will behave. That's the problem that Cambridge, MA-based startup Perceptive Automata is tackling head-on with a unique approach to deep learning that builds on research into human perception. Co-founder and CTO Sam Anthony joins the show to explain how he's teaching self-driving cars how to tap into he power of the human glance.