“Photonics Shines in the Age of AI”
👉Check out SOITEC Silicon Photonics white paper in the March 2024 issue of Photonics Spectra magazine. The issue covers how photonics is and will be shaping future scaling, cost, and performance of the AI infrastructure!
Link to the white paper 👉 https://lnkd.in/dVYpNsMh
Engineered silicon substrates are providing the foundation for the cutting-edge photonics engines that data centers will need to usher in the era of AI.
#semiconductor#SOI#technology#AI
“Photonics Shines in the Age of AI”
👉Check out SOITEC Silicon Photonics white paper in the March 2024 issue of Photonics Spectra magazine. The issue covers how photonics is and will be shaping future scaling, cost, and performance of the AI infrastructure!
Link to the white paper 👉 https://lnkd.in/dVYpNsMh
We are delighted to announce the publication of our latest work on hardware-efficient photonic tensor cores for AI acceleration in Nanophotonics! In this work, we designed a scalable and efficient optical dot-product engine using customized multi-operand photonic devices, namely multi-operand optical neurons (MOONs). We taped out a multi-operand-Mach–Zehnder-interferometer (MOMZI) as a starting point and experimentally demonstrated its utility in image recognition tasks. By utilizing MOON-based photonic tensor cores, we were able to significantly reduce the device footprint and improve the hardware efficiency of implementing tensor operations in the optical domain. Thanks to our collaborators Jiaqi Gu, Hanqing Zhu, Shupeng Ning, Rongxing Tang, May Hlaing, Jason Midkiff, Sourabh Jain, and our supervisors, prof. David Z. Pan and prof. Ray Chen. Check out the paper online at the provided link!
https://lnkd.in/gaAH4gF5#AIacceleration#integratedphotonics#siliconphotonics#opticalcomputing#deeplearning
This new hybrid electronic-optical model significantly outperforms systems that only utilize either a diffractive optical network or an electronic neural network for optical information transfer through diffusive random media, highlighting the importance of having both an electronic encoder and a diffractive decoder that work together. #Optics#Photonics#FreeSpaceOptics
Our most recent paper, titled "Physical Human-Robot Interaction Control of an Upper Limb Exoskeleton with a Decentralized Neuroadaptive Control Scheme," has been published as early access in the IEEE Transactions on Control Systems Technology journal. In this paper:
1) A human upper limb model and an exoskeleton model are decentralized and augmented at the subsystem level to enable decentralized control action design,
2) The estimation of human exogenous torque (HET) using radial basis function neural networks (RBFNNs) along with human arm and robot inertial parameters enabled the designed controller to be adaptable to different operators,
3) Barrier Lyapunov function (BLF) is incorporated for the first time into the virtual decomposition control scheme to ensure the safety of the operator,
4) Unknown actuator constraints and unknown model uncertainties in a commercial haptic exoskeleton of Haption, ABLE, are tackled by utilizing RBFNNs to increase the robustness of the controller,
5) The asymptotic stability of the entire system and robustness of the controller are proved and validated by performing extensive experiments in the presence of unknown disturbances, different human operators, and different working velocities.
For those interested in further details, the full article can be accessed through the following link:
https://lnkd.in/diBJWgQK/#pHRI#nonlinear_control#impedance_Control#radial_basic_function_neural_networks#haptics#exoskeletons
📢 We are excited to introduce our final Focused Session at EOSAM 2024: FS5 - Machine-Learning for Optics and Photonic Computing for AI
In the past few years, artificial intelligence (AI) techniques have opened up new horizons for research into photonics. AI techniques are particularly powerful to design photonics structures and devices for specific tasks, often resulting in higher efficiency and generally improved performance. Photonics components and systems are also exceedingly considered to replace conventional electronic implementations of AI computation. This session aims to cover the most recent findings and scientific insight in advanced techniques and applications of this merger between Photonics and AI.
Invited speakers in this Focused Session:
- Benjamin Wetzel: Characterization and machine-learning optimization of modulation instability processes in nonlinear fiber optics
- Jasper Riebesehl: Machine learning techniques for noise characterization of optical frequency combs
More information: https://lnkd.in/df_Tfr7D
Ensure your participation by registering for the conference today!
🔗 Registration Link: https://lnkd.in/dGTACket#EOSAM2024#MachineLearning#Optics#Photonics#Computing#AI#Conference#Naples
In this session from our #NTT R&D Forum 2023, Bob Byer from Stanford University and NTT RESEARCH, and Timothy McKenna from NTT Research, expanded on the topic of Lithium Niobate Photonics in the era of #AI, including episodes involving breakthroughs from their long history and tradition.
Watch it here: https://lnkd.in/gvMGfBVP
JUST IN: Researchers at the University of Illinois Urbana-Champaign have developed an AI breakthrough enhancing Atomic Force Microscopy resolution. This revolutionary technique promises to transform nanoelectronics and material studies. #AI#AFM#Innovation#Nanoelectronics#TheReportify
Let us know your thoughts in the comments!
To learn more and read the complete story, click on this link.
https://lnkd.in/gaZg9gea
𝐍𝐞𝐰 𝐩𝐮𝐛𝐥𝐢𝐜𝐚𝐭𝐢𝐨𝐧 𝐨𝐟 𝐭𝐡𝐞 𝐂𝐨𝐦𝐩𝐮𝐭𝐚𝐭𝐢𝐨𝐧𝐚𝐥 𝐎𝐩𝐭𝐢𝐜𝐬 𝐠𝐫𝐨𝐮𝐩 𝐚𝐭 𝐂𝐡𝐚𝐢𝐫 𝐓𝐎𝐒!
We are pleased to announce the publication of our latest research paper titled "High Fidelity Laser Beam Shaping Using Liquid Crystal on Silicon Spatial Light Modulators as Diffractive Neural Networks." Since laser beam shaping with LCoS-SLMs often suffers from beam shaping artifacts, partly caused by unconsidered properties of the LCoS devices, this publication presents a method to consider and compensate for these inherent properties of LCoS devices by treating the SLM as a diffractive neural network.
🔍 𝐖𝐡𝐚𝐭'𝐬 𝐍𝐞𝐱𝐭?
In next steps it is planned to implement more complex models and automatically fit the corresponding parameters from experimental results. This will enable the method to be easily applicable to arbitrary SLMs without having to determine the parameters manually. It is also planned to analyze the effects and their correction on DNNs with multiple SLMs.
Access the full publication here: https://lnkd.in/eeR6m-mj
(Co-)Authors: Paul Buske, Oskar Hofmann, Annika Bonhoff, Jochen Stollenwerk, and Prof. Dr. Carlo Holly Holly
RWTH Aachen University / Fraunhofer ILT#WirAmTOS#tosrwth#tosrwthaachen#rwthaachen#research#EUVTechnology#EUV#OpticalSystems#research
