People, Ideas, and the Path Ahead
summary
Describes a claim that answer set programming plus a disciplined rigorous KR methodology is practical: an ASP-based RCS/USA-Advisor for the space shuttle produces plans and diagnostics in seconds (many cases <2s, all under 14s) versus 20 minutes for human experts. Shows modular reuse (digital-circuits module lifted unchanged), develops EasyCSP (ASP as host with constraint solvers) and deploys it for Kodak scheduling with incremental rescheduling, and implements time-fluents plus sparse timelines and clingo-based incremental solving for a Super Bowl COVID access-control system handling 30–40k people.
AI-written summary · about
chapters (9)
- 0:00Start
- 2:59Michael Gelfond
- 22:20Facing a Practical Problem
- 26:06Hybrid Declarative Languages
- 36:00David Ferrucci
- 39:01IBM Watson Meets NMR
- 43:07Lessons Learned from Covid
- 46:08Still a Challenge
- 49:33Incremental Computation
related talks
38meXtreme Programming with Artificial Intelligence
30mFrom Constraints to Cognition: A Hybrid Framework for Adaptive, Explainable(…)
46mPanel: AI in the Loop - Left-Shifting Platform DX and Governance
28mBalancing Human Oversight and AI in UXR
38mAI for Software Development: A Reality Check
20mBuild Systems, Not Code
13mChallenges, Considerations, and Best Practices
50mFrom Prompt Hacking to Architectural Determinism: Engineering Reliable GenAI Systems