Susan L. Epstein

The CUNY Graduate School, Department of Computer Science and

 Hunter College, Department of Computer Science

     Home 

     Publications 

     Collaborators 

     Courses 

     Media Appearances 

     Contact

FORR

FORR is a satisficing architecture for learning and problem solving. Its premise is that agreement among reactivity, heuristics, planning, and search is a valid decision- making principle. A FORR-based program learns to specialize general domain expertise into problem-class specific expertise. An implemented FORR shell supports domain-specific development. FORR explores the relationship among reactivity, heuristics, and search. Applications include Hoyle for game playing, Ariadne for robot path finding and ACE for constraint satisfaction. Each of them has demonstrated substantial, learned expertise after relatively little training.

Current Work

- The integration of planning with high-level reasoning.

- Learning new heuristics.

- Autonomous restructuring of a decision hierarchy.

- Metaknowledge for heuristics.

- Multiagent solutions to multiple goal problems with (CD).

Adaptive Constraint Engine (ACE)

Game playing (Hoyle)

Wayfinding (Ariadne)

Human-Multi-robot Teams (SemaFORR)

Spoken Dialogue Systems (FORRSooth)

Key references

Epstein, S. L., & Petrovic, S. (2012). Learning Expertise with Bounded Rationality and Self-awareness. In Metareasoning: Thinking about thinking: MIT Press.

Epstein, S. L., E. C. Freuder and M. Wallace 2005. Learning to Support Constraint Programmers. Computational Intelligence. 21(4): 337-371.

Epstein, S. L. 1994. For the Right Reasons: The FORR Architecture for Learning in a Skill Domain. Cognitive Science, 18 (3): 479-511.