New Frontiers for Information Processing & Computing: Attojoule Nanophotonics and Photonic Algorithms

Date(s):

Location:
EBU1 Booker 2512

Sponsored By:
Zhaowei Liu

Speaker(s):
Volker J. Sorger
Volker J. Sorger

Abstract:

In this talk I will discuss my team’s latest research combining device physics with computer engineering and show intersections with information theory. We are inspired to find answers to both fundamental and technological questions in the fields of optoelectronics, computing, information science. Here, our vision is to execute a mathematical algorithm or a compute-instruction by flowing light through a reconfigurable system where data enters and information exits. As an example, we designed a temporal FFT capable of handling terabits-per-second of data at a fraction of the energy of GPUs. Furthermore, we are interested exploring novel hardware compute-engines from first principles beyond von-Neumann architectures; for instance, our research in the field of neuromorphic computing uses electrooptic nonlinearity as an activation function enabling vector matrix multiplications at the speed of light in neural networks with applications in deep-learning, nonlinear optimization, and image pre-processing. Furthermore, using the coincidence detection property of a leaky-integrate-and-fire neuromorphic network, I will share our latest results in revealing the symmetry of an image – a property key for human 3D vision and even argued to be a signature for intelligence. Beyond cognitive networks, we are intrigued by reconfigurable photonic accelerators for PDE solvers using one-shot execution parallelism such as in residue number-system arithmetic. Lastly, we are interested in finding revolutionary ways to solve NP-complete ‘hard’ decision problems, whose exponentially exponential time complexity functions could be dramatically reduced by implementing graph theory problems using nanoscale building blocks. Particular topics discussed in this talk include a) fundamental scaling laws for optoelectronics, b) demonstrations of the first 100aJ/bit efficient and 60GHz-fast modulators, c) quantum tunnel light sources, e) photonic routers enabling RNS compute engines, f) photonic FFTs and NP-complete oracles, and g) neuromorphic photonic computing.


Speaker Bio:
Volker J. Sorger is an associate professor in the Department of Electrical and Computer Engineering, and the director of the Orthogonal Physics Enabled Nanophotonics (OPEN) lab at the George Washington University. He received his PhD from the University of California Berkeley. His research areas include optoelectronics, plasmonics and nanophotonics, including novel materials, and optical information processing and neuromorphic computing. Amongst his breakthroughs are the first demonstration of a semiconductor plasmon laser, a 100aJ/bit electrooptic modulator, optical FFT on-chip, and symmetry-detection, neuromorphic nanophotonic computing. Dr. Sorger received multiple awards among are the Early Career Award and Dean’s Outstanding Young Research Award at GW, the AFOSR Young Investigator Award, Hegarty Innovation Prize, and the National Academy of Sciences award of the year. Dr. Sorger is the editor-in-chief of Nanophotonics, the OSA chair for the Optoelectronics and Photonics division, and serves at the board of meetings at OSA and SPIE and the scholarship committee. He is a senior member of IEEE, OSA & SPIE.

Contact:
Valeria Rendon
vrendon@ucsd.edu
858-822-6939