**Probabilistic Learning of Human Driving Behavior for Creating Realistic Simulations**
6:00p Welcome reception
6:30p Presentation with Q&A
Predicting human driving behavior has long been recognized as one of the key capabilities of an autonomous vehicle that’s used to facilitate planning. More recently, there has been a lot of interest in using human behavior models to construct realistic simulations, where autonomous vehicles can be trained and tested safely and efficiently. While in many respects similar to on-board prediction, modeling human behavior in a simulation context presents unique problems and opportunities. In this talk I will describe how those considerations inform our technology development at Inverted AI, where we produce simulated traffic that’s realistic, diverse, and reactive.