Title :
A poly-time analysis of robustness in feedforward neural networks
Author :
Alippi, Cesare ; Moioli, Milena
Author_Institution :
Dipartimento di Elettronica e Inf., Politecnico di Milano, Italy
Abstract :
The paper provides a methodology for evaluating the performance degradation of a feedforward neural network once affected by fixed perturbations injected in the computation. The loss in performance associated with the perturbed computation can be evaluated with a polynomial time algorithm and a general family of loss functions by resorting to a probabilistic analysis based on randomized algorithms. The methodology can be used to test the impact of finite precision implementations (analog, digital or mixed) on the weights of a feedforward neural network
Keywords :
computational complexity; feedforward neural nets; performance evaluation; randomised algorithms; feedforward neural networks; finite precision implementations; fixed perturbations; performance degradation; polynomial time algorithm; probabilistic analysis; randomized algorithms; robustness; Algorithm design and analysis; Computer networks; Degradation; Feedforward neural networks; Neural networks; Performance analysis; Performance loss; Polynomials; Robustness; Testing;
Conference_Titel :
Virtual and Intelligent Measurement Systems, 2001, IEEE International Workshop on. VIMS 2001
Conference_Location :
Budapest
Print_ISBN :
0-7803-6568-2
DOI :
10.1109/VIMS.2001.924905