Building fast Bayesian computing machines out of intentionally stochastic, digital parts.

作者: Eric Jonas , Vikash K. Mansinghka

DOI:

关键词:

摘要: The brain interprets ambiguous sensory information faster and more reliably than modern computers, using neurons that are slower less reliable logic gates. But Bayesian inference, which underpins many computational models of perception cognition, appears computationally challenging even given transistor speeds energy budgets. principles structures needed to narrow this gap unknown. Here we show how build fast computing machines intentionally stochastic, digital parts, narrowing efficiency by multiple orders magnitude. We find connecting stochastic components according simple mathematical rules, one can massively parallel, low precision circuits solve inference problems compatible with the Poisson firing statistics cortical neurons. evaluate for depth motion perception, perceptual learning causal reasoning, each performing over 10,000+ latent variables in real time - a 1,000x speed advantage commodity microprocessors. These results suggest new role randomness engineering reverse-engineering intelligent computation.

参考文章(40)
Joshua B. Tenenbaum, Vikash Kumar Mansinghka, Natively probabilistic computation Massachusetts Institute of Technology. ,(2009)
John R. Anderson, Michael Matessa, A rational analysis of categorization international conference on machine learning. pp. 76- 84 ,(1990) , 10.1016/B978-1-55860-141-3.50013-4
Michael B. Elowitz, Stanislas Leibler, A synthetic oscillatory network of transcriptional regulators Nature. ,vol. 403, pp. 335- 338 ,(2000) , 10.1038/35002125
C Rosenmund, J. Clements, G. Westbrook, Nonuniform probability of glutamate release at a hippocampal synapse Science. ,vol. 262, pp. 754- 757 ,(1993) , 10.1126/SCIENCE.7901909
Persi Diaconis, The Markov chain Monte Carlo revolution Bulletin of the American Mathematical Society. ,vol. 46, pp. 179- 205 ,(2008) , 10.1090/S0273-0979-08-01238-X
H. H. McAdams, A. Arkin, Stochastic mechanisms in gene expression Proceedings of the National Academy of Sciences of the United States of America. ,vol. 94, pp. 814- 819 ,(1997) , 10.1073/PNAS.94.3.814
Aaron P Blaisdell, Kosuke Sawa, Kenneth J Leising, Michael R Waldmann, Causal reasoning in rats. Science. ,vol. 311, pp. 1020- 1022 ,(2006) , 10.1126/SCIENCE.1121872
Konrad P Körding, Daniel M Wolpert, None, Bayesian integration in sensorimotor learning Nature. ,vol. 427, pp. 244- 247 ,(2004) , 10.1038/NATURE02169
Paul Dagum, Michael Luby, An optimal approximation algorithm for Bayesian inference Artificial Intelligence. ,vol. 93, pp. 1- 27 ,(1997) , 10.1016/S0004-3702(97)00013-1
Mingjie Lin, Ilia Lebedev, John Wawrzynek, High-throughput bayesian computing machine with reconfigurable hardware field programmable gate arrays. pp. 73- 82 ,(2010) , 10.1145/1723112.1723127