Reinforcement-learning AIs are vulnerable to a new kind of attack

Mar 2, 2020

Reinforcement-learning AIs are vulnerable to a new kind of attack

Posted by Genevieve Klien in categories: robotics/AI, transportation

Adversarial attacks against the technique that powers game-playing AIs and could control self-driving cars shows it may be less robust than we thought.

Leave a reply

GETAS THREAT LEVEL: GUARDED
TOP DONOR: BRIAN CARTMELL
#2 DONOR: DONALD MACLEAN
#3 DONOR: JEFF BONE/TOPSIGHT FUND
#4 DONOR: PEG KAY

LF FUND: $594,573.68
BITCOIN ENDOWMENT: $7,320,935.83
A-PRIZE: $30,013.00
NANOSHIELD: $17,018.00
X-RISKS NETWORK: $3,108.00
AISHIELD: $4,112.00

LIFEBOAT NEWS: 3,022 SUBSCRIBERS
GETAS ALERTS: 436 SUBSCRIBERS
BLOG: 48,721 POSTS
BOARDS: 2,844 MEMBERS
REPORTS: 55
PROGRAMS: 25
FORUMS: 13
QUOTES: 136

FACEBOOK: 8,341 MEMBERS
LINKEDIN: 1,983 MEMBERS
TWITTER FEED: 33,526 MEMBERS
FIGHT AIDS: 3 MEMBERS
[email protected]: 15 MEMBERS
[email protected]: 44 MEMBERS
A-PRIZE: 55 MEMBERS
NANOSHIELD: 69 MEMBERS
AISHIELD: 25 MEMBERS

DONORS: 3,923
LF-500: 13 MEMBERS
PATRON: 1 MEMBER
FIRST CLASS: 1 MEMBER
EXECUTIVE: 12 MEMBERS
ECONOMY: 23 MEMBERS
A-500: 1 MEMBER

© 2002–2020 Lifeboat Foundation