- Abstract:
-
Using the achievements of my research group over the last 30+ years, I provide evidence to support two hypotheses:
1. The representation of a problem is often the key to its solution;
2. Problem representations can be automatically both formed and repaired.
I focus mainly on mathematical problem solving based on logical representations, drawing on work in solving mechanics problems in English, assisting ecologists to construct simulation models, reasoning with diagrams, providing an analogy-based functional program editor and repairing faulty mathematical theories and ontologies. I look especially at the interaction between reasoning and representation, illustrating how failures in reasoning can suggest representational improvements and repairs. Methodologically, interactive systems can provide a vehicle for focusing on some aspects of representation construction, while relying on user interaction to complement automation.
- Copyright:
- 2006 by Dartmouth College, USA. All Rights Reserved
- Links To Paper
- No links available
- Bibtex format
- @InProceedings{EDI-INF-RR-0841,
- author = {
Alan Bundy
},
- title = {Constructing, Selecting and Repairing Representations of Knowledge},
- book title = {Proceedings of AI@50},
- year = 2006,
- month = {Jul},
- }
|