Big Problems Posed as Challenges to AI Researchers
What constitutes an AI Grand Challenge?
Here's what Ron Brachman proposed in his July 2005 AAAI Presidential Address:(AA)AI - More than the Sum of its Parts (AI Magazine 27(4): Winter 2006, 19-34):
Grand Challenge Criteria [slide #41]
- Clear and compelling demonstration of cognition
- Non-gameable proxy for a range of problems requiring cognitive capabilities
- Clear and simple measurement
- Success is well-defined
- Decomposable and diagnostic
- Partial results; failure should point to way to needed improvement
- Ambitious and visionary, but not unrealistic
- Faith in success in 10 - 20 years
- Compelling to the general public
- Motivating for the research community
Image from DARPA Robotics Challenge.