A5062306355- Photonic and Optical Devices
- Optical Network Technologies
- Neural Networks and Reservoir Computing
- Advanced Photonic Communication Systems
- Semiconductor Lasers and Optical Devices
- Phase-change materials and chalcogenides
- Advanced Fiber Laser Technologies
- Photonic Crystals and Applications
- Mechanical and Optical Resonators
- Plasmonic and Surface Plasmon Research
- Advanced Biosensing Techniques and Applications
- Metamaterials and Metasurfaces Applications
- Advanced Memory and Neural Computing
- Advanced Optical Sensing Technologies
- Nanowire Synthesis and Applications
- Optical and Acousto-Optic Technologies
- Random lasers and scattering media
- Nanofabrication and Lithography Techniques
- Semiconductor materials and devices
- Optical Coherence Tomography Applications
- 3D IC and TSV technologies
- bioluminescence and chemiluminescence research
- Neural dynamics and brain function
- Maritime Navigation and Safety
- Underwater Vehicles and Communication Systems
University of Florida
2023-2024
George Washington University
2022-2024
Institute of Electrical and Electronics Engineers
2023
Politecnico di Milano
2013-2022
Consorzio di Bioingegneria e Informatica Medica
2016
Photonic Random-Access Memories (P-RAM) are an essential component for the on-chip non-von Neumann photonic computing by eliminating optoelectronic conversion losses in data links. Emerging Phase-Change Materials (PCMs) have been showed multilevel memory capability, but demonstrations still yield relatively high optical loss and require cumbersome WRITE-ERASE approaches increasing power consumption system package challenges. Here we demonstrate a multistate electrically programmed low-loss...
The field of mimicking the structure brain on a chip is experiencing interest driven by demand for machine intelligent applications. However, power consumption and available performance machine-learning (ML) accelerating hardware still leave much desire improvement. In this letter, we share viewpoints, challenges, prospects electronic-photonic neural network (NN) accelerators. Combining electronics with photonics offers synergistic co-design strategies high-performance AI...
Solving mathematical equations faster and more efficiently has been a Holy Grail for centuries scientists engineers across all disciplines. While electronic digital circuits have revolutionized equation solving in recent decades, it become apparent that performance gains from brute-force approaches of compute-solvers are quickly saturating over time. Instead, paradigms leverage the universes' natural tendency to minimize system's free energy, such as annealers or Ising Machines, being sought...
Metalenses are emerging as an alternative to digital micromirror devices (DMDs), with the advantages of compactness and flexibility. The exploration metalenses has ignited enthusiasm among optical engineers, positioning them forthcoming frontier in technology. In this paper, we advocate for implementation phase-change material, Sb2Se3, capable providing swift, reversible, non-volatile focusing defocusing within 1550 nm telecom spectrum. lens, equipped a robust ITO microheater, offers...
Abstract The retention-of-state functionality provided by memories is fundamental to any Turing machine and neural network, hence critical for information system today. While emerging optical learning accelerators photonic neuromorphic computing paradigms provide promising signal processing performance, the lack of a photon-photon force in universe makes storing challenging. Fortunately, phase change materials such missing memristive nonvolatile function via their reconfigurable crystalline...
Key to recent successes in the field of artificial intelligence (AI) has been ability train a growing number parameters which form fixed connectivity matrices between layers nonlinear nodes. This "deep learning" approach AI historically required an exponential growth processing power far exceeds computational throughput digital hardware as well trends efficiency. New computing paradigms are therefore enable efficient information while drastically improving throughput. Emerging strategies for...
This roadmap consolidates recent advances while exploring emerging applications, reflecting the remarkable diversity of hardware platforms, neuromorphic concepts, and implementation philosophies reported in field. It emphasizes critical role cross-disciplinary collaboration this rapidly evolving
A photonic integrated circuit performing simultaneous mode and wavelength demultiplexing for few-mode-fiber transmission is demonstrated the first time. The realized on an InP-based technological platform; it can handle up to eight mode- wavelength-division-multiplexed (MDM/WDM) channels allows all-optical multiple-input-multiple-output processing unscramble mixing generated by fiber propagation. single arrayed waveguide grating used demultiplex WDM carried all propagating modes, optimizing...
Abstract Machine learning and artificial intelligence (AI) is becoming a ubiquitous technology of the looming industry-4.0 era. However, progress adopting intelligent automation systems limited by hardware overhead such as throughput, power consumption, latency. At conceptual level, electronics at end its scaling law alternative accelerators are sought after. Optical co-processors offer high-degree algorithmic homomorphism to implement general matrix-matrix multiplication operations via...
Here we demonstrate a photonic tensor core based on silicon photonics dot-product engine. Utilizing compact electronic phase-change-material memory and WDM show the highest throughput density to date of 3.8 MAC/s/mm 2 .
Solving mathematical equations faster and more efficiently has been a Holy Grail for centuries scientists engineers across all disciplines. While electronic digital circuits have revolutionized equation solving in recent decades, it become apparent that performance gains from brute-force approaches of compute-solvers are quickly saturating over time. Instead, paradigms leverage the universes natural tendency to minimize systems free energy, such as annealers or Ising Machines, being sought...
We present our implementation of a Convolution Neural Network employing an Integrated Photonic Chip (PIC) to perform the high-speed optical FFT, showing Silicon design, initial response, and packaging.
This study introduces a Silicon Photonic Tensor Core, exemplifying advanced integration and harnessing optical computing’s speed energy efficiency. Demonstrating image convolution with less than 7% error, the system showcases substantial potential for computing applications.
We demonstrate on-chip non-invasive monitoring of orthogonal modes transmitted in a silicon photonic waveguide. The proposed technique exploits recently developed ContactLess Integrated Photonic Probe (CLIPP) realizing fully transparent integrated light detector. optical intensity the propagating waveguide is tracked time by CLIPP, with no signal quality degradation induced operations. exploit this concept for simultaneous two modulated 10 Gbit/s data channels at same wavelength and...
In this paper, we present an improved version of Application-Specific Photonic Integrated Chip for solving Partial Differential Equations (PDEs). This novel chip is designed to solve PDEs with specific reflecting boundary conditions, by means a series integrated mirrors at the edges.
We demonstrate a 4-channel silicon photonic MIMO demultiplexer performing all-optical unscrambling of four mixed modes. Real-time on-chip light monitoring through transparent detectors enables robust demultiplexing 10 Gbit/s channels with less than −20 dB crosstalk.
As Internet-of-Things (IoT) devices continue to grow rapidly in number, developing energy-efficient memory solutions has become critically important. This paper introduces an innovative Phase Change Memory (PCM) architecture that can significantly reduce energy consumption IoT devices. After highlighting the energy-inefficiency of current designs, we explore possibilities leveraging PCM. We demonstrate benefits exploiting PCM are dependent on working frequency CPU and show how surpass with...