Put AI to work
June 26-27, 2017: Training
June 27-29, 2017: Tutorials & Conference
New York, NY

Inverse reward design

Anca Dragan (UC Berkeley)
11:55am12:35pm Thursday, June 29, 2017
Location: Sutton South/Regent Parlor
Average rating: *****
(5.00, 3 ratings)

What you'll learn

  • Understand why AI agents should have uncertainty about their objectives and use human input as valuable observations to improve their estimates

Description

As AI agents become more capable of optimizing their objective functions, it’s becoming increasingly important to make sure that we give them the right objectives in the first place. Unfortunately, humans are notoriously bad at specifying what we want. Anca Dragan explains why agents should have uncertainty about their objectives and use human input as valuable observations to improve their estimates.

Photo of Anca Dragan

Anca Dragan

UC Berkeley

Anca Dragan is an assistant professor in the EECS Department at UC Berkeley. Her goal is to enable robots to work with, around, and in support of people. Anca runs the InterACT Lab, where she focuses on algorithms for human-robot interaction—algorithms that move beyond the robot’s function in isolation and generate robot behavior that also accounts for interaction and coordination with end users. The lab works across different applications, from assistive robots to manufacturing to autonomous cars, and draws from optimal control, planning, estimation, learning, and cognitive science. Anca also helped found and serves on the steering committee for the Berkeley AI Research (BAIR) Lab and is a co-PI of the Center for Human-Compatible AI.