We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Physics > Optics

Title: Generative Deep Learning Model for a Multi-level Nano-Optic Broadband Power Splitter

Abstract: We propose a novel Conditional Variational Autoencoder (CVAE) model, enhanced with adversarial censoring and active learning, for the generation of 550 nm broad bandwidth (1250 nm to 1800 nm) power splitters with arbitrary splitting ratio. The device footprint is 2.25 x 2.25 {\mu} m2 with a 20 x 20 etched hole combination. It is the first demonstration to apply the CVAE model and the adversarial censoring for the photonics problems. We confirm that the optimized device has an overall performance close to 90% across all bandwidths from 1250 nm to 1800 nm. To the best of our knowledge, this is the smallest broadband power splitter with arbitrary ratio.
Subjects: Optics (physics.optics); Signal Processing (eess.SP); Computational Physics (physics.comp-ph)
Cite as: arXiv:2003.03747 [physics.optics]
  (or arXiv:2003.03747v1 [physics.optics] for this version)

Submission history

From: Keisuke Kojima [view email]
[v1] Sun, 8 Mar 2020 08:59:53 GMT (1586kb,D)

Link back to: arXiv, form interface, contact.