MIT makes autonomous driving decisions easier to interpret
· By Antonio Sedino, CTRO · Published by Reinventy Solutions Corp.
MIT describes CW-Net, a method that translates an autonomous vehicle's AI reasoning into understandable concepts.
The research aims to help people anticipate errors. It addresses interpretation of system behaviour, rather than claiming that an autonomous vehicle will no longer make mistakes.
Read the original source at MIT News — Artificial Intelligence ↗