Title :
A cost function for backward chaining inference
Author :
Karkishchenko, A.N.
Author_Institution :
Lab. of Math. Problems of Artificial Intell., Taganrog Technol. Inst. of Southern Fed. Univ., Taganrog
Abstract :
A notion of a cost function for acquiring generalized solutions in logical inference systems is introduced. An example of such systems is widely used backward chaining inference machine. It is shown that the cost function is monotone and sub additive. It is also offered a method of a so called rank approximation of the cost function that reduces the computational complexity of its applying.
Keywords :
approximation theory; computational complexity; graph theory; inference mechanisms; backward chaining inference; computational complexity; cost function; logical inference systems; rank approximation; Artificial intelligence; Boring; Computational complexity; Cost function; Expert systems; Hydrogen; Laboratories; Lattices; Logic programming; Production systems;
Conference_Titel :
Fuzzy Information Processing Society, 2008. NAFIPS 2008. Annual Meeting of the North American
Conference_Location :
New York City, NY
Print_ISBN :
978-1-4244-2351-4
Electronic_ISBN :
978-1-4244-2352-1
DOI :
10.1109/NAFIPS.2008.4531228
