Stochastic Mixed-Signal VLSI Architecture for High-Dimensional Kernel Machines

Abstract

A mixed-signal paradigm is presented for high-resolution parallel inner- product computation in very high dimensions, suitable for efficient im- plementation of kernels in image processing. At the core of the externally digital architecture is a high-density, low-power analog array performing binary-binary partial matrix-vector multiplication. Full digital resolution is maintained even with low-resolution analog-to-digital conversion, ow- ing to random statistics in the analog summation of binary products. A random modulation scheme produces near-Bernoulli statistics even for highly correlated inputs. The approach is validated with real image data, and with experimental results from a CID/DRAM analog array prototype in 0.5

Cite

Text

Genov and Cauwenberghs. "Stochastic Mixed-Signal VLSI Architecture for High-Dimensional Kernel Machines." Neural Information Processing Systems, 2001.

Markdown

[Genov and Cauwenberghs. "Stochastic Mixed-Signal VLSI Architecture for High-Dimensional Kernel Machines." Neural Information Processing Systems, 2001.](https://mlanthology.org/neurips/2001/genov2001neurips-stochastic/)

BibTeX

@inproceedings{genov2001neurips-stochastic,
  title     = {{Stochastic Mixed-Signal VLSI Architecture for High-Dimensional Kernel Machines}},
  author    = {Genov, Roman and Cauwenberghs, Gert},
  booktitle = {Neural Information Processing Systems},
  year      = {2001},
  pages     = {1099-1105},
  url       = {https://mlanthology.org/neurips/2001/genov2001neurips-stochastic/}
}