TR2020-025

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


Abstract:

A novel Conditional Variational Autoencoder (CVAE) model with the adversarial censoring is presented to help to generate the 550nm broad bandwidth (1250nm to 1800nm) power splitter with arbitrary splitting ratio

 

  • Related News & Events

    •  NEWS    MERL Scientists Presented 5 Papers Including 2 Invited Talks at Optical Fiber Communications Conference (OFC) 2020
      Date: March 8, 2020 - March 13, 2020
      MERL Contacts: Devesh K. Jha; Toshiaki Koike-Akino; Kieran Parsons; Ye Wang
      Research Areas: Communications, Electronic and Photonic Devices, Machine Learning, Signal Processing
      Brief
      • Due to COVID-19, MERL Optical Team scientists remotely presented 5 papers including 2 invited talks at the Optical Fiber Communications Conference (OFC) 2020, that was held in San Diego from March 8-13, 2020. Topics presented include recent advances in quantum signal processing, channel coding design, nano-optic power splitter, and deep learning-based integrated photonics. In addition, Dr. Kojima gave an invited workshop talk on deep learning-based nano-photonic device optimization.

        OFC is the largest global conference and exhibition for optical communications and networking professionals. The program is comprehensive from research to marketplace, from components to systems and networks and from technical sessions to the exhibition. For over 40 years, OFC has drawn attendees from all corners of the globe to meet and greet, teach and learn, make connections and move the industry forward. The five-day technical conference features peer reviewed presentations and more than 180 invited speakers, the thought leaders in the industry presenting the highlights of emerging technologies. Additional technical programming throughout the week includes special symposia, special sessions, in-depth tutorials, workshops, panels and the thought-provoking rump session.
    •