ACDC: A Structured Efficient Linear Layer · HackerPapers discussions

ACDC: A Structured Efficient Linear Layer — mentioned by bnegreve

Sat, 26 Jan 2019 09:09:08 GMT

There are Moczulski et al. with their ACDC paper [1]. They describe a structured layer that can approximate any linear transform and discuss an optical implementation of it (see section 1.1 "lightning fast deep sell" :). Our deep SELL offers several possibilities for analog physical implementation. Given the great demand for fast low energy neural networks, the possibility of harnessing physical phenomena to perform efficient computation in deep networks is worthy of consideration.... (see paper) [1] https://arxiv.org/pdf/1511.05946.pdf

Paper and discussion history. A mention is not an endorsement.

ACDC: A Structured Efficient Linear Layer — mentioned by stochastic_monk

Mon, 12 Feb 2018 15:50:26 GMT

This reminds me of ACDC, a Deep-fried Convnets-like[0] approach by some NVIDIA employees. [1] See section 1.1, where they state they could perform the operation in analog. [0]: https://arxiv.org/abs/1412.7149 [1]: https://arxiv.org/pdf/1511.05946

Paper and discussion history. A mention is not an endorsement.

ACDC: A Structured Efficient Linear Layer — mentioned by dharma1

Fri, 26 Aug 2016 09:49:04 GMT

This paper describes an optical implementation to some transforms used in deep learning http://arxiv.org/abs/1511.05946 These guys use optical computing for big data analysis and pattern matching http://optalysys.com/

Paper and discussion history. A mention is not an endorsement.