Photonic circuits and probabilistic computing

Placeholder Show Content

Abstract/Contents

Abstract
This work explores computing in the noisy, stochastic, low-power setting, particularly with photonic systems. Stochasticity and noisy operation has been with computing since its earliest days, before the development of extremely reliable components (e.g. CMOS) enabled reliable digital and analog operation and relegated noise to a low level of abstraction. As the limits of current technologies are reached, as power consumption becomes a bottleneck for computing, and as novel stochastic algorithms are developed, a probabilistic approach extends usefully beyond the level of single components: from stochastic encoding of information to entire circuit architectures. Photonic systems - which can involve the interaction of a handful of photons with a handful of internal degrees of freedom like atomic states - provide a natural platform for consideration of noisy, low-power computing. This is a work in two parts. In Part I I investigate applications of photonic circuits for power-efficient computation. I propose a circuit architecture for the decoding of low density parity-check codes that uses only optical waveguides, noisy optical switches, and is inherently tolerant of faults affecting its components; one of its most appealing features is the graceful degradation of performance and computation time as we lower the power in the laser that powers the device. I also introduce various design motifs that I hope will be useful to future optical engineers. Part II has a more information theoretic flavor. I consider several optics-inspired models for encoding information stochastically - e.g., in the distribution of optical power over multiple waveguides, or multiple frequency bands for a Gaussian channel, prove some optimality results about some of the schemes I propose, and discuss energy-efficient communication for these setups.

Description

Type of resource text
Form electronic; electronic resource; remote
Extent 1 online resource.
Publication date 2014
Issuance monographic
Language English

Creators/Contributors

Associated with Pavlichin, Dmitri Serguei
Associated with Stanford University, Department of Physics.
Primary advisor Mabuchi, Hideo
Thesis advisor Mabuchi, Hideo
Thesis advisor Doniach, S
Thesis advisor Weissman, Tsachy
Advisor Doniach, S
Advisor Weissman, Tsachy

Subjects

Genre Theses

Bibliographic information

Statement of responsibility Dmitri Serguei Pavlichin.
Note Submitted to the Department of Physics.
Thesis Thesis (Ph.D.)--Stanford University, 2014.
Location electronic resource

Access conditions

Copyright
© 2014 by Dmitri Pavlichin
License
This work is licensed under a Creative Commons Attribution 3.0 Unported license (CC BY).

Also listed in

Loading usage metrics...