Modern humanity and the global economy rely on optical infrastructures and technologies to drive our digital society. We will continue to see exponential data traffic. The cost of this growth in communications draws a significant fraction of the energy society uses. New, more energy-efficient approaches and technologies must be developed and adapted to permit continued information growth. Our information and communications lifeline demands innovation and game-changing discoveries to keep up.
The 2022 winners of the Optica Foundation Challenge in the information category will present their progress along with a special keynote.
The optical implementation of neural networks can be advantageous compared to electronics in terms of power consumption. This derives from the fact that the energy required to transmit information optically can be nearly independent of the distance between the emitter and the receiver. Consequently, optics can be particularly suitable for hardware implementations of neural networks due to the dense connectivity of neural architectures. Neural networks, however, use the strengths of the interconnections between the processing units (the “neurons”) as computing and storage elements. An optical neural network must therefore include a mechanism that allows it to be programmed or trained by modifying the strength of the interconnections. In this presentation we will review optical methods and also present recent results for programming optical learning machines implemented with multi-mode fibers.
 Teğin, U., Yıldırım, M., Oğuz, İ, C. Moser, D. Psaltis, Scalable optical learning operator. Nat Comput Sci 1, 542–549 (2021).
Compiling Deep Learning Tasks onto (Quantum-) Optical Systems
The hardware limitations of conventional electronics in deep neural network (DNN) applications have spurred exploration into alternative architectures, including optical accelerators. This work investigates the scalability and performance metrics—such as throughput, energy consumption, and latency—of various optical and opto-electronic architectures, with a focus on recently developed hardware error correction techniques, in-situ training methods, initial field trials, as well as extensions into DNN-based inference on quantum signals with reversible, quantum-coherent resources.
Mengjie Yu, University of Southern California, USA
Mark Lawrence, Washington University in St. Louis, USA
Chaoran Huang, Chinese University of Hong Kong, Hong Kong
15:30 - 15:35
Alan Willner, University of Southern California, USA
Chair, Optica Foundation Challenge Selection Committee
Chaoran Huang, Chinese University of Hong Kong, Hong Kong (2022 Challenge Winner)
Talk Title: Integrated photonic neuromorphic processor enables intelligent, energy-efficient signal processing for the next-generation communication systems
Mark Lawrence, Washington University in St. Louis, USA (2022 Challenge Winner)
Talk Title: Fast, low-power, and high-resolution meta-reflect-arrays for massive space-division-multiplexing
Dirk Englund, MIT, USA
Talk Title: Compiling Deep Learning Tasks onto (Quantum-) Optical Systems
View the Environment Symposium l View the Health Symposium