Probabilistic logic networks: a comprehensive framework for uncertain inference

This book describes Probabilistic Logic Networks (PLN), a novel conceptual, mathematical and computational approach to uncertain inference. In order to carry out effective reasoning in real-world circumstances, AI software must robustly handle uncertainty. However, previous approaches to uncertain inference donot have the breadth of scope required to provide an integrated treatment of the disparate forms of cognitively critical uncertainty as they manifest themselves within the various forms of pragmatic inference. Going beyond prior probabilistic approaches to uncertain inference, PLN is able to encompass within uncertain logic such ideas as induction, abduction, analogy, fuzziness and speculation, and reasoning about time and causality. The book reviews the conceptual and mathematical foundations of PLN, giving the specific algebra involved in each type of inference encompassed within PLN. Inference control and the integration of inference with other cognitive faculties are also briefly discussed. Provides a comprehensive framework for uncertain reasoning, integrating probability theory, predicate and term logic, and pattern theory Considers a broad scope of reasoning types Fuses rigorous mathematics with practical computation to describe methods designed for large-scale and, in many cases, real-time inference within commercial software systems INDICE: Introduction.- Knowledge Representation.- Experiential Semantics.-Indefinite Truth Values.- First-Order Extensional Inference: Rules and Strength Formulas.- First-Order Extensional Inference with Indefinite Truth Values.-First-Order Extensional Inference with Distributional Truth Values.- Error Magnification in Inference Formulas.- Holistic Inference Strategies.- Extensional Inference on Predicates.- Handling Crisp and Fuzzy Quantifiers with Indefinite Truth Values.- Intensional Inference.- Inference Control.- Temporal and Causal Inference.- Appendix: Comparison of PLN Rules and NARS Rules.- Bibliography.- Index.

• ISBN: 978-0-387-76871-7
• Editorial: Springer
• Encuadernacion: Cartoné
• Páginas: 300
• Fecha Publicación: 01/06/2008
• Nº Volúmenes: 1
• Idioma: Inglés