Papers by Lorenzo De Marinis
Neural Computing and Applications, Apr 25, 2022
Photonics-based neural networks promise to outperform electronic counterparts, accelerating neura... more Photonics-based neural networks promise to outperform electronic counterparts, accelerating neural network computations while reducing power consumption and footprint. However, these solutions suffer from physical layer constraints arising from the underlying analog photonic hardware, impacting the resolution of computations (in terms of effective number of bits), requiring the use of positive-valued inputs, and imposing limitations in the fan-in and in the size of convolutional kernels. To abstract these constraints, in this paper we introduce the concept of Photonic-Aware Neural Network (PANN) architectures, i.e., deep neural network models aware of the photonic hardware constraints. Then, we devise PANN training schemes resorting to quantization strategies aimed to obtain the required neural network parameters in the fixed-point domain, compliant with the limited resolution of the underlying hardware. We finally carry out extensive simulations exploiting PANNs in image classification tasks on well-known datasets (MNIST, Fashion-MNIST, and Cifar-10) with varying bitwidths (i.e., 2, 4, and 6 bits). We consider two kernel sizes and two pooling schemes for each PANN model, exploiting 2 Â 2 and 3 Â 3 convolutional kernels, and max and average pooling, the latter more amenable to an optical implementation. 3 Â 3 kernels perform better than 2 Â 2 counterparts, while max and average pooling provide comparable results, with the latter performing better on MNIST and Cifar-10. The accuracy degradation due to the photonic hardware constraints is quite limited, especially on MNIST and Fashion-MNIST, demonstrating the feasibility of PANN approaches on computer vision tasks.
Photonic-Aware Neural Network: a fixed-point emulation of photonic hardware
2022 27th OptoElectronics and Communications Conference (OECC) and 2022 International Conference on Photonics in Switching and Computing (PSC), Jul 3, 2022
Photonic-aware Neural Networks for Packet Classification in Beyond 5G Networks
A Lithium Niobate on Insulator Based Photonic Neural Network
2022 27th OptoElectronics and Communications Conference (OECC) and 2022 International Conference on Photonics in Switching and Computing (PSC), Jul 3, 2022
Silicon Photonic Filter-based Dot Product Engine for Convolutional Neural Networks
OSA Advanced Photonics Congress 2021, 2021
We present a silicon photonic filter-based analog engine for computing dot products in convolutio... more We present a silicon photonic filter-based analog engine for computing dot products in convolutional neural networks. It shows a greater energy efficiency compared to electronic solutions with a limited bit resolution degradation of input signals.
In the modern era of artificial intelligence, increasingly sophisticated artificial neural networ... more In the modern era of artificial intelligence, increasingly sophisticated artificial neural networks (ANNs) are implemented, which pose challenges in terms of execution speed and power consumption. To tackle this problem, recent research on reduced-precision ANNs opened the possibility to exploit analog hardware for neuromorphic acceleration. In this scenario, photonic-electronic engines are emerging as a short-medium term solution to exploit the high speed and inherent parallelism of optics for linear computations needed in ANN, while resorting to electronic circuitry for signal conditioning and memory storage. In this paper we introduce a precision-scalable integrated Photonic-Electronic Multiply-Accumulate Neuron (PEMAN). The proposed device relies on (i) an analog photonic engine to perform reduced-precision multiplications at high speed and low power, and (ii) an electronic front-end for accumulation and application of the nonlinear activation function by means of a nonlinear encoding in the analog-to-digital converter (ADC). The device has been numerically validated through cosimulations to perform multiply-accumulate operations (MAC). Simulations are based on the iSiPP50G SOI process for the photonic engine and a commercial 28 nm CMOS process for the electronic front-end. The PEMAN exhibits a multiplication accuracy of 6.1 ENOB up to 10 GMAC/s, while it can perform computations up to 56 GMAC/s with a reduced accuracy down to 2.1 ENOB. The device can trade off speed and power consumption with resolution, significantly outperforming its analog electronics counterparts both in terms of speed and energy consumption. With respect to other photonic ANNs, the PEMAN has comparable speed and energy consumption with a higher resolution, while outperforming them by a hundredfold in the fan-in, which opens the possibility to accelerate more complex networks. Index Terms-Photonic-electronic codesign, photonic neural networks, photonic analog computing, neural network accelerator, reduced precision computing. I. INTRODUCTION N OWADAYS machine learning technology is pervasively used for a wide range of applications including image
The characterization of a broadband Si 3 N 4 integrated linear optical processor operating in the... more The characterization of a broadband Si 3 N 4 integrated linear optical processor operating in the C-band is reported. The impact of losses on the processor accuracy is discussed towards the photonic implementation of state-of-the-art neural networks.
A programmable P4 node performing neural network acceleration is demonstrated. The node implement... more A programmable P4 node performing neural network acceleration is demonstrated. The node implements both traffic feature extraction and neural networking computations. Cyber security use case targeting DDoS attack detection is deployed.
Applied sciences, Jul 5, 2021
This article is an open access article distributed under the terms and conditions of the Creative... more This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY
Improving Noise Resilience in End-to-End Deep Learning Optical Fiber Transmission Links
OSA Advanced Photonics Congress (AP) 2020 (IPR, NP, NOMA, Networks, PVLED, PSC, SPPCom, SOF), 2020
The characterization of a broadband low-loss 4 × 4 MZI-based reconfigurable linear optical proces... more The characterization of a broadband low-loss 4 × 4 MZI-based reconfigurable linear optical processor is reported. The impact of MZI extinction ratio on the effective number of bits (ENOB) at the device output is also investigated.
A Photonic Accelerator for Feature Map Generation in Convolutional Neural Networks
OSA Advanced Photonics Congress (AP) 2020 (IPR, NP, NOMA, Networks, PVLED, PSC, SPPCom, SOF), 2020
We propose a photonic accelerator for convolutional neural networks carrying out linear, nonlinea... more We propose a photonic accelerator for convolutional neural networks carrying out linear, nonlinear, and pooling operations. It achieves an excellent accuracy in TensorFlow simulations and is more energy efficient than state-of-the-art electronics.
Physics-inspired End-to-End Deep Learning for High-Performance Optical Fiber Transmission Links
We experimentally demonstrate the performance improvements obtained through End-to-End Deep Learn... more We experimentally demonstrate the performance improvements obtained through End-to-End Deep Learning in noise and chromatic dispersion compensation of optical fiber transmission links when incorporating a physics-inspired activation function compared to state-of-the-art ReLU configurations.
Photonic integrated neural network accelerators
The paper reports MZI-based photonic integrated devices fabricated in silicon- on-insulator and s... more The paper reports MZI-based photonic integrated devices fabricated in silicon- on-insulator and silicon nitride platforms acting as linear optical processors, and discusses their suitability for accelerating state-of-the-art neural networks for computer vision.
Advancing Legal and Practical Recognition of the Non-Human Right to Energy
Routledge eBooks, Dec 12, 2022
Photonic-aware Neural Networks for Packet Classification in URLLC scenarios
Telemetry and AI-based security P4 applications for optical networks [Invited]
Journal of Optical Communications and Networking, Sep 29, 2022
This paper presents the potentials and challenges of programmable packet-optical nodes encompassi... more This paper presents the potentials and challenges of programmable packet-optical nodes encompassing coherent pluggable modules applied in the context of optical metro networks. Two innovative applications of P4-based data plane programmability are then presented. The first targets the monitoring and processing of optical telemetry data/metadata directly in the forwarding plane. The second one focuses on the deployment of deep neural networks in P4 chipsets, effectively supporting in-network distributed cyber security functionalities in packet-optical nodes.
IEEE Access, 2019
Photonic solutions are today a mature industrial reality concerning high speed, high throughput d... more Photonic solutions are today a mature industrial reality concerning high speed, high throughput data communication and switching infrastructures. It is still a matter of investigation to what extent photonics will play a role in next-generation computing architectures. In particular, due to the recent outstanding achievements of artificial neural networks, there is a big interest in trying to improve their speed and energy efficiency by exploiting photonic-based hardware instead of electronic-based hardware. In this work we review the state-of-the-art of photonic artificial neural networks. We propose a taxonomy of the existing solutions (categorized into multilayer perceptrons, convolutional neural networks, spiking neural networks, and reservoir computing) with emphasis on proof-of-concept implementations. We also survey the specific approaches developed for training photonic neural networks. Finally we discuss the open challenges and highlight the most promising future research directions in this field. INDEX TERMS Artificial neural networks, neural network hardware, photonics, neuromorphic computing, photonic neural networks.
IEEE Journal of Selected Topics in Quantum Electronics, 2023
While the use of graphic processing units fueled the success of artificial intelligence models, t... more While the use of graphic processing units fueled the success of artificial intelligence models, their future evolution will likely require overcoming the speed and energy efficiency limitations of current implementations with the use of specialized neuromorphic hardware. In this scenario, neuromorphic photonic processors have recently proved to be a feasible solution. In this paper, we first discuss basic analog photonic processing elements based on Mach-Zehnder modulators and assess their effective bit resolution. Then we evaluate how different photonic integration technologies affect the performance and the scalability of analog optical processors, in order to provide a clearer path toward real-world implementations of such engines. To this aim, we focus our analysis on the silicon on insulator (SOI), lithium niobate on insulator (LNOI), and indium phosphide (InP) platforms. In particular, we have numerically evaluated the performance of the Photonic Electronic Multiply-Accumulate Neuron (PEMAN) and its tensorial version, both based on Mach-Zehnder modulators, with the three technologies in terms of resolution, energy efficiency, and footprint efficiency. LNOI modulators achieve the best resolution at high speed, with 4.3 bits at 56 GMAC/s for the single PEMAN and 3.6 bits at 224 GMAC/S for the tensorial version. The energy consumption in InP and LNOI platforms is the lowest, accounting for just 13.2 pJ/MAC and 4.6 pJ/MAC for the single and tensorial PEMAN, respectively. Nonetheless, SOI devices outperform the others in terms of footprint efficiency, reaching 18.6 GMAC/s/mm 2 in the single-neuron version and 29.6 GMAC/s/mm 2 in the tensorial version. Index Terms-photonic analog computing, photonic neural networks, photonic integration technologies.
High-performance end-to-end deep learning IM/DD link using optics-informed neural networks
Optics Express, May 31, 2023
In this paper, we introduce optics-informed Neural Networks and demonstrate experimentally how th... more In this paper, we introduce optics-informed Neural Networks and demonstrate experimentally how they can improve performance of End-to-End deep learning models for IM/DD optical transmission links. Optics-informed or optics-inspired NNs are defined as the type of DL models that rely on linear and/or nonlinear building blocks whose mathematical description stems directly from the respective response of photonic devices, drawing their mathematical framework from neuromorphic photonic hardware developments and properly adapting their DL training algorithms. We investigate the application of an optics-inspired activation function that can be obtained by a semiconductor-based nonlinear optical module and is a variant of the logistic sigmoid, referred to as the Photonic Sigmoid, in End-to-End Deep Learning configurations for fiber communication links. Compared to state-of-the-art ReLU-based configurations used in End-to-End DL fiber link demonstrations, optics-informed models based on the Photonic Sigmoid show improved noise- and chromatic dispersion compensation properties in fiber-optic IM/DD links. An extensive simulation and experimental analysis revealed significant performance benefits for the Photonic Sigmoid NNs that can reach below BER HD FEC limit for fiber lengths up to 42 km, at an effective bit transmission rate of 48 Gb/s.
Uploads
Papers by Lorenzo De Marinis