We introduce an algorithm that learns gradients from samples in the supervised learning framework. An error analysis is given for the convergence of the gradient estimated by the ...
Signal strength fluctuation is one of the major problems in a fingerprint-based localization system. To alleviate this problem, we propose a scrambling method to exploit temporal d...
It has been observed by E. Eberlein and U. Keller that the hyperbolic distribution fits logarithmic rates of returns of a stock much better than the normal distribution. We give a...
In this paper, performance formulae for a queue serving Gaussian traffic are presented. The main technique employed is motivated by a general form of Schilder's theorem, the ...
— Parallel algorithms are presented for modules of learning automata with the objective of improving their speed of convergence without compromising accuracy. A general procedure...