Abstract
This chapter discusses adaptive decision elements. The general objective of adaption in vote-takers is to give the reliable inputs more influence in determining the output than the unreliable inputs have. The chapter describes the systematic procedures, each of which could be directly carried out by electrical circuits that could be used to obtain an estimate of the best vote-weight. The simplest adaption procedure, called adaption method I, is a procedure in which the error probability of the output of a circuit is inferred from conditions in the circuit, such as a bias current. The quantity that indicates the error probability of the output of the circuit is used directly to set the vote-weight of the output in all later vote-takers. This is an open-loop adaption procedure, so it is stable and requires no special analysis. In a cyclic error-counting adaption procedure, the vote-weights are changed periodically, based on data collected during the computations in the period. Because cyclic error-counting adaption procedures are based on a fixed number of equally weighted observations, their analysis is sufficiently tractable to give quantitative insights into the behavior of error-counting adaptive systems.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call