Theoretical Computer Science
Tractable reasoning via approximation
Artificial Intelligence
Epistemological and heuristic adequacy revisited
ACM Computing Surveys (CSUR)
A nonstandard approach to the logical omniscience problem
Artificial Intelligence
Experimental results on the crossover point in random 3-SAT
Artificial Intelligence - Special volume on frontiers in problem solving: phase transitions and complexity
A Computing Procedure for Quantification Theory
Journal of the ACM (JACM)
A computational model of belief
Artificial Intelligence
A machine program for theorem-proving
Communications of the ACM
A Deduction Model of Belief
Annals of Mathematics and Artificial Intelligence
Implementing the Davis–Putnam Method
Journal of Automated Reasoning
TARK '94 Proceedings of the 5th conference on Theoretical aspects of reasoning about knowledge
Does This Set of Clauses Overlap with at Least One MUS?
CADE-22 Proceedings of the 22nd International Conference on Automated Deduction
| Hi-index | 0.00 |
It is well-known that the logicist approach to agency is confronted with both epistemological and heuristic problems. On the one hand, the agent's model must be logically adequate: it must provide us a clear picture of what the agent is, and is not, able to deduce from its background knowledge. On the other hand, the agent's program must be adequate in practice: it must generate useful conclusions from input data and given the computational resources that are actually available. In actuality, the agent's need for heuristic adequacy has strong epistemological consequences. Based on this argument, this paper proposes a framework which is based on the paradigm of knowledge approximation and that is flexible enough to incorporate heuristic strategies used in satisfiability algorithms. The framework is used as a "logical toolbox" for modelling resource-bounded agents that have different operational means at their disposal to approximate knowledge. The toolbox consists in a family of relative relevance logics which are semantically founded on the notion of resource and that include interesting features, such as incremental reasoning and dual approximations.