Jury-trial Story Construction and Analysis using Goal-directed Answer Set Progra...
summary
Goal-directed Answer Set Programming with s(CASP) top-down ASP models Commonwealth of Massachusetts v. Johnson as facts and defaults, then queries for verdicts; it generates justification trees that explain why a branch yields murder, self-defense or hung jury and supports abductive and counterfactual hypotheses. The system avoids grounding (reducing blowup), but runtime grows with the number of abductive assumptions and alternatives; demonstrations successfully produce multiple consistent verdicts.
AI-written summary · about
chapters (13)
- 0:00Jury trial automation project
- 1:30Pennington and Hastie’s model
- 4:43Answer Set Programming basics
- 8:38Goal-directed ASP execution
- 10:01Trial knowledge modeling
- 16:27Mental models and emotions
- 18:56Verdicts and self-defense
- 21:49Counterfactual limits
- 22:52Conclusion and applications
- 23:15Q&A on logic
- 23:56Q&A on linguistic aspects and modeling
- 24:38Q&A on ChatGPT
- 26:28Live ASP solver demo
related talks
13mBlawx: User-friendly Goal-Directed Answer Set Programming for Rules as Code
30mFrom Constraints to Cognition: A Hybrid Framework for Adaptive, Explainable(…)
56mPeople, Ideas, and the Path Ahead
17mHow to Program like a Five Year Old in Haskell
29mMarketplace Logistic via Answer Set Programming
39mMaking AI Reliable in the SDLC: The Role of Deterministic Context at Scale
30mToward Programming Languages for Reasoning -- Humans, Symbolic Systems, and AI ...
01h 03mImproved Assistance for Interactive Proof