Seeding Algorithm for Bipolar Stochastic Computing for Polynomial Approximations

Application of stochastic computing (SC) for reckoning trascendental functions as tanh(x), so used as an activation function in convolutional neural networks is an active research area. Currently, most of the works for computing functions via SC are based on the unipolar encoding format (x 2 0; 1), due to this, a method based on bipolar encoding format (x 2 -1; 1) is here proposed with the goal of reducing the implementation complexity, and the correlation between stochastic bitstreams. For that, a collection of existing methods is adapted for the purpose of this work. Moreover, for reducing correlation between bitstream, an algorithm is proposed for the selection of different seeds for distinct linear feedback shift registers that yields to a lowMSE. The seed selection along with the adaptation of methods for implementing polynomials with SC digital circuits based on a bipolar encoding format yields to more accurate results. Simulations were carried out for the polynomial approximation of several functions. Function tanh(x) was compared with an existing solution, verifying in that way the superior performance of the proposed approach.