Picture of Mathieu Chagnon

Mathieu Chagnon

Stuttgart, Germany
MTS, Research Scientist


  • Post-Doc, Nokia Bell Labs, Stuttgart, Germany & University of Stuttgart, Institute of Telecommunications, Germany. 2017
  • Ph.D., Electrical & Computer Engineering, McGill University. 2016.
    Title: Optical communications for long-haul, short-reach, and chip-scale distances enabled by digital signal processing.
    Link: https://escholarship.mcgill.ca/concern/theses/9p290d31m
  • Master of Engineering, Electrical & Computer Engineering, McGill University. 2011.
    Title: Digital signal processing assessment for optical coherent receiver using dual-polarization Quadrature phase shift keying modulation.
    Link: https://escholarship.mcgill.ca/concern/theses/3484zh17v
  • Bachelor of Science, Engineering Physics, Polytechnique Montréal. 2008.


Mathieu Chagnon is a Member of Technical Staff and an optical transmission systems researcher in the Smart Optical Fabric and Devices Lab at Nokia Bell Labs. He has coauthored more than 100 peer-reviewed publications including several invited and tutorial papers, a book chapter on high-speed interconnects, and patents in the field of fiber-optic communication systems. He has given invited and tutorial talks at major conference venues in China, Europe, and the USA. His research interests include Digital Signal Processing and Machine Learning for communication systems, multi-dimensional modulation and signal recovery, as well as information theory. More specific to machine learning, Mathieu won the first prize of both editions of the Nokia AI Conference demonstrating methods for digital communication systems to self-optimize and learn to communicate. The Nokia AI Conference is in an internal event where all of Nokia’s 100’000 employees are invited to submit proposals and compete for the most impactful and innovative solution leveraging AI. Moreover, in collaboration with colleagues from several Nokia Bell Labs locations, he also demonstrated the first quantization and implementation of a deep neural network running in real time on an FPGA extracting the soft-decision bits directly from a photocurrent at the receiving end of a fiber-optic communication system. Dr. Chagnon is on the Editorial Board of the IEEE Communications Magazine.

Research Interests

  • Artificial Intelligence
  • Computational & Algorithmic Sciences
  • Deep Learning
  • Machine Learning
  • Optical Communication

Honors and Awards

  • NSERC Alexander Graham Bell Canada Graduate Scholarship
  • Postdoctoral Fellowship from the Natural Sciences and Engineering Research Council of Canada
  • Les Vadasz Doctoral Fellowships in Engineering
  • SPIE Optics and Photonics Graduate Scholarship
  • IEEE Photonics Society Graduate Student Fellowship
  • Frist prize 2018 Nokia AI Conference
  • Frist prize 2019 Nokia AI Conference


Selected Articles and Publications

  • M. Chagnon, J. Siirtola, T. Rissa, A. Verma, “Quantized Deep Neural Network Empowering an IM–DD Link Running in real time on a Field Programable Gate Array”, in Proc. of European Conference on Optical Communication, 2019.
  • B. Karanov, M. Chagnon, F. Thouin, T. A. Eriksson, H. Bulow, D. Lavery, P. Bayvel, and L. Schmalen, “End-to-End Deep Learning of Optical Fiber Communications,” in Journal of Lightwave Technology, vol. 36, no. 20, 2018
  • M. Chagnon, “Optical Communications for Short Reach,” (Tutorial) in Journal of Lightwave Technology, vol. 37, no. 8, 2019
  • M. Chagnon, M. Morsy-Osman, D. V. Plant, “Half-Terabit Single-Carrier Direct-Detect Transceiver, Formats, and DSP: Analysis and Demonstration,” in Journal of Lightwave Technology, vol. 36, no. 2, 2018


Books and Chapters

  • D.V. Plant, M.H. Morsy-Osman, M. Chagnon, S. Lessard (2018) “Trends in High-Speed Interconnects for Datacenter Networking: Multidimensional Formats and Their Enabling DSP”. In: Optical Switching in Next Generation Data Centers. Springer, Cham, 2017


  1. Optical transmitters and receivers using polarization multiplexing, Patent US10116410B2
  2. Probabilistic signal shaping and forward error correction using subcarrier multiplexing, Patent EP3474470A1