Reproduce “A Survey on Silicon Photonics for Deep Learning” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Febin P Sunny, Ebadollah Taheri, Mahdi Nikdast, Sudeep Pasricha
The authors survey photonic accelerators, mapping devices, neurons, and network architectures. They explain Broadcast-and-Weight: wavelengths carry signals in parallel, microring resonators set weights, and photodetectors sum signals. A hitless design is described as reducing thermal crosstalk, but its microring count is unreported. Crosslight reported 9.5× lower energy per bit and 15.9× higher performance than two photonic accelerators, not electronic systems.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.