Accurate and compact stochastic computations by exploiting correlation

Abstract

Recent studies have shown, contrary to what was previously believed, that by exploiting correlation instochastic computing (SC) designs, more accurate SC circuits with low area cost can be realized. However, if these basicSC circuits or blocks are cascaded in series to form a large complex system, correlation between stochastic numbers(SNs) from one block to the next would be lost; thus, inaccuracies are introduced. In this study, we propose correlatingcircuits to be used in building complex correlated SC systems. One of the circuits is the correlator that restores lostcorrelations between two SNs due to previous processing. In addition, a correlated SN generator is introduced togenerate SN correlated to a specific SN. Experimental results show that our methods have improved the accuracy ofstochastic computation and preserved the stochastic computing correlation without the need for conversion from SCto the conventional binary-encoded computing, and vice versa. Consequently, lower latency and lower area cost areachieved.