We are proud to showcase the results of research undertaken by members of our interdisciplinary Artificial Intelligence, Photonics. Our work and findings have been published in a number of prestigious Journals in our field.
GRAPHENE OPTICAL MODULATOR WITH INTEGRATED ULTRA-HIGH PERFORMANCE
As data processing, transmission, and storage needs expand, it's harder to optimize optoelectronic devices' high-speed and energy usage Silicon- and nitride-based heterogeneous photonics have high-speed promise but millimeter- to centimeter-scale footprints. The hunt for a high-speed, energy-efficient, small electro-optic modulator continues. We accomplish 60 GHz speed (3 dB roll-off) with a double-layer graphene optical modulator on a Silicon photonics platform. Vertical distributed-Bragg-reflector cavity enhances electro-optic responsiveness 40-fold while maintaining modulation depth (5.2 dB/V). Small, efficient, rapid modulators provide high photonic chip density and performance for signal processing, sensor platforms, and analog- and neuromorphic photonic processors.
BREAKING THE POWER AND SPEED LIMIT OF LASERS: NEW FAST, POWERFUL COMPACT LASER INVENTED
OP has developed a new generation of vertical-cavity surface-emitting laser (VCSEL) that demonstrates record-fast temporal bandwidth. This was possible by combining multiple transverse coupled cavities, which enhances optical feedback of the laser. VCSELs have emerged as a vital approach for realizing energy-efficient and high-speed optical interconnects in data centers and supercomputers.
STRAINOPTRONICS: A NEW WAY TO CONTROL PHOTONS
We have discovered a new way to engineer optoelectronic devices by stretching a two-dimensional material on top of a silicon photonic platform. The amount of strain these semiconductor 2D materials can bear is significantly higher when compared to bulk materials for a given amount of strain. They also note these novel 2D material-based photodetectors are 1,000 times more sensitive compared to other photodetectors using graphene. Photodetectors capable of such extreme sensitivity are useful not only for data communication applications but also for medical sensing and possibly even quantum information systems. said Dr. Volker J. Sorger, Founder and President of Optelligence LLC.
HETEROGENEOUSLY INTEGRATED ITO PLASMONIC MACH-ZEHNDER
INTERFEROMETRIC MODULATOR ON SOI
By heterogeneously adding a thin material layer of indium tin oxide to the silicon photonic waveguide chip, we have demonstrated an optical index change 1,000 times larger than silicon. Unlike many designs based on resonators, this spectrally broadband device is stable against temperature changes and allows a single fiber-optic cable to carry multiple wavelengths of light, increasing the amount of data that can move through a system.
“We are delighted to have achieved this decade-long goal of demonstrating a GHz-fast ITO modulator. This sets a new horizon for next-generation photonic reconfigurable devices with enhanced performance yet reduced size,” said Dr. Volker J. Sorger, Founder and President of Optelligence LLC.
BREAKING THE SIZE AND SPEED LIMIT OF MODULATORS FOR NEXT GENERATION INTERNET AND COMMUNICATION NETWORKS
Unlike the current paradigm in electronic machine learning hardware that processes information sequentially, this processor uses the Fourier optics, a concept of frequency filtering which allows for performing the required convolutions of the neural network as much simpler element-wise multiplications using the digital mirror technology
"This massively parallel amplitude-only Fourier optical processor is heralding a new era for information processing and machine learning. We show that training this neural network can account for the lack of phase information," said Dr. Volker J. Sorger, Founder and President of Optelligence LLC.
"This prototype demonstration shows a commercial path for optical accelerators ready for a number of applications like network-edge processing, data-centers, and high-performance compute systems," said Dr. Hamed Dalir, Co-founder and CTO of Optelligence LLC.
DIAGNOSTIC BREAST CANCER USING CONVEX FACTORIZATION EMBEDDING THERMOGRAPHY
Thermography detects breast cancer early (CBE). Computer thermography matrix factorization can extract thermal variations. These methods highlight thermographic sequences. Change patterns are hard to capture in infrared. This study depicts the whole sequence in one image using convex factor analysis and bell-curve membership function embedding. Thermomics were extracted from this low-dimensional (LD) temperature sequence to train a breast cancer diagnosis algorithm. Comparing embedding and factorization. Combining clinical and demographic data produces 78.9% (75.7%, 85.9%); convex-NMF alone yields 76.9% (73.7%, 86.1%). The proposed approach preserves thermal patterns, improving CBE and early breast cancer diagnosis.
AI MACHINES CAN LEARN UNSUPERVISED AT SPEED OF LIGHT
“The performance of the light neural network (tensor) processing units TPU was found to be three orders higher than a regular electrical TPU, with photons also found to be an ideal match for computing node-distributed networks as well as engines performing advanced tasks with high throughput at the edge of a network, like 5G.” said Dr. Volker J. Sorger, Founder and President of Optelligence LLC.