STOCHASTIC AND DETERMINISTIC COMPUTING: DIGITAL FILTERING AND OTHER APPLICATIONS.
Date of Original Version
Two approaches to performing arithmetic operations with 'one-bit' words and their application to digital filtering are discussed. In the first approach the numbers, scaled to lie between 0 and 1, are represented by probabilities. Then the similarity between probability algebra and Boolean algebra is exploited to obtain simple digital filter implementations. This approach, known as stochastic computing, inherently has inaccuracies associated with it, since the arithmetic result is a probability and has to be estimated. However, this approach is resistant to soft errors. In the second approach the 'run' properties of pseudorandom sequences are used to implement digital filters deterministically using exact arithmetic.
Conference Record - Asilomar Conference on Circuits, Systems & Computers
Gupta, P. K., and R. Kumaresan. "STOCHASTIC AND DETERMINISTIC COMPUTING: DIGITAL FILTERING AND OTHER APPLICATIONS.." Conference Record - Asilomar Conference on Circuits, Systems & Computers , (1987): 159-162. https://digitalcommons.uri.edu/ele_facpubs/718