Title :
The complexity of recurrent neural computations
Author :
Shumsky, Sergey A.
Author_Institution :
P.N. Lebedev Phys. Inst., Acad. of Sci., Moscow, Russia
Abstract :
The paper considers the amount of information processed in recurrent neural networks with random interconnections. The complexity of computations is defined as the number of neurons switchings during the run. The variety of neurons in the network gives rise to distribution of computational complexity. The author calculates this complexity distribution, and finds the limits of complexity in the individual algorithms
Keywords :
computational complexity; recurrent neural nets; complexity distribution; computational complexity; neuron switchings; random interconnections; recurrent neural computations; Computational complexity; Computer architecture; Computer networks; Convergence; Distributed computing; Feedback loop; Integrated circuit interconnections; Neurons; Recurrent neural networks; Switches;
Conference_Titel :
Neuroinformatics and Neurocomputers, 1995., Second International Symposium on
Conference_Location :
Rostov on Don
Print_ISBN :
0-7803-2512-5
DOI :
10.1109/ISNINC.1995.480862
