Title :
Non-suspiciousness: a generalisation of convexity in the frame of foundations of numerical analysis and learning
Author :
Bianchini, M. ; Fanelli, S. ; Gori, M. ; Protasi, M.
Author_Institution :
Dept. of Syst. & Inf., Florence Univ., Italy
Abstract :
The effectiveness of connectionist models in emulating intelligent behaviour is strictly related to the capability of the learning algorithms to find optimal or near-optimal solutions. In this paper, a canonical reduction of gradient descent dynamics is proposed, allowing the formulation of the neural network learning as a finite continuous optimisation problem, under some nonsuspiciousness conditions. In the linear case, the nonsuspect nature of the problem guarantees the implementation of an iterative method with O(n2) as computational complexity. Finally, since nonsuspiciousness is a generalisation of the concept of convexity, it is possible to apply this theory to the resolution of nonlinear problems
Keywords :
computational complexity; conjugate gradient methods; iterative methods; learning (artificial intelligence); neural nets; optimisation; canonical reduction; computational complexity; connectionist models; convexity generalisation; finite continuous optimisation problem; gradient descent dynamics; intelligent behaviour; iterative method; learning; near-optimal solutions; neural network learning; nonlinear problems; nonsuspiciousness; numerical analysis; Books; Computer networks; Convergence; Humans; IEEE members; Iterative methods; Machine learning; Neural networks; Numerical analysis; Shape;
Conference_Titel :
Neural Networks Proceedings, 1998. IEEE World Congress on Computational Intelligence. The 1998 IEEE International Joint Conference on
Conference_Location :
Anchorage, AK
Print_ISBN :
0-7803-4859-1
DOI :
10.1109/IJCNN.1998.686020
