On Robustness of Sparse High-D Computing
Konferenz: CNNA 2016 - 15th International Workshop on Cellular Nanoscale Networks and their Applications
23.08.2016 - 25.08.2016 in Dresden, Deutschland
Tagungsband: CNNA 2016
Seiten: 2Sprache: EnglischTyp: PDF
Persönliche VDE-Mitglieder erhalten auf diesen Artikel 10% Rabatt
Autoren:
Laiho, Mika; Poikonen, Jussi H.; Lehtonen, Eero (Technology Research Center, University of Turku, Finland)
Kanerva, Pentti (Redwood Center for Theoretical Neuroscience, University of California, Berkeley, USA)
Inhalt:
In this paper we discuss how High-Dimensional (High-D) computing tolerates different nonidealities in the computing hardware. In High-D computing, information is distributed holistically into vectors of high dimensionality, so that all bits carry an equal amount of information. The computing is inherently statistical, as random High-D vectors are combined by vector-arithmetic operations. Here the effect of selected nonidealities to robustness is studied with simulations.