Combinatorial optimization problems play crucial roles in real-world applications, and many studies from a physics perspective have contributed to specialized hardware for high-speed computation. However, some combinatorial optimization problems are easy to solve, and others are not. Hence, the qualification of the difficulty in problem-solving will be beneficial. In this paper, we employ the Koopman analysis for multiple time-series data from the replica exchange Monte Carlo method. After proposing a quantity that aggregates the information of the multiple time-series data, we performed numerical experiments. The results indicate a negative correlation between the proposed quantity and the ability of the solution search.
{"title":"Koopman analysis of combinatorial optimization problems with replica exchange Monte Carlo method","authors":"Tatsuya Naoi, Tatsuya Kishimoto, Jun Ohkubo","doi":"arxiv-2409.03154","DOIUrl":"https://doi.org/arxiv-2409.03154","url":null,"abstract":"Combinatorial optimization problems play crucial roles in real-world\u0000applications, and many studies from a physics perspective have contributed to\u0000specialized hardware for high-speed computation. However, some combinatorial\u0000optimization problems are easy to solve, and others are not. Hence, the\u0000qualification of the difficulty in problem-solving will be beneficial. In this\u0000paper, we employ the Koopman analysis for multiple time-series data from the\u0000replica exchange Monte Carlo method. After proposing a quantity that aggregates\u0000the information of the multiple time-series data, we performed numerical\u0000experiments. The results indicate a negative correlation between the proposed\u0000quantity and the ability of the solution search.","PeriodicalId":501083,"journal":{"name":"arXiv - PHYS - Applied Physics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142177802","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Guillermo Nava Antonio, Quentin Remy, Jun-Xiao Lin, Yann Le Guen, Dominik Hamara, Jude Compton-Stewart, Joseph Barker, Thomas Hauet, Michel Hehn, Stéphane Mangin, Chiara Ciccarelli
The optical manipulation of magnetism enabled by rare earth-transition metal ferrimagnets holds the promise of ultrafast, energy efficient spintronic technologies. This work investigates laser-induced picosecond spin currents generated by ferrimagnetic GdCo via terahertz emission spectroscopy. A suppression of the THz emission and spin current is observed at magnetization compensation when varying the temperature or alloy composition in the presence of a magnetic field. It is demonstrated that this is due to the formation of domains in the GdCo equilibrium magnetic configuration. Without an applied magnetic field, the picosecond spin current persists at the compensation point. The experimental findings support the model for THz spin current generation based on transport of hot spin-polarized electrons, which is dominated by the Co sublattice at room temperature. Only at low temperature a comparable contribution from Gd is detected but with slower dynamics. Finally, spectral analysis reveals a blueshift of the THz emission related to the formation of magnetic domains close to magnetization compensation.
{"title":"Origin of the laser-induced picosecond spin current across magnetization compensation in ferrimagnetic GdCo","authors":"Guillermo Nava Antonio, Quentin Remy, Jun-Xiao Lin, Yann Le Guen, Dominik Hamara, Jude Compton-Stewart, Joseph Barker, Thomas Hauet, Michel Hehn, Stéphane Mangin, Chiara Ciccarelli","doi":"arxiv-2409.03088","DOIUrl":"https://doi.org/arxiv-2409.03088","url":null,"abstract":"The optical manipulation of magnetism enabled by rare earth-transition metal\u0000ferrimagnets holds the promise of ultrafast, energy efficient spintronic\u0000technologies. This work investigates laser-induced picosecond spin currents\u0000generated by ferrimagnetic GdCo via terahertz emission spectroscopy. A\u0000suppression of the THz emission and spin current is observed at magnetization\u0000compensation when varying the temperature or alloy composition in the presence\u0000of a magnetic field. It is demonstrated that this is due to the formation of\u0000domains in the GdCo equilibrium magnetic configuration. Without an applied\u0000magnetic field, the picosecond spin current persists at the compensation point.\u0000The experimental findings support the model for THz spin current generation\u0000based on transport of hot spin-polarized electrons, which is dominated by the\u0000Co sublattice at room temperature. Only at low temperature a comparable\u0000contribution from Gd is detected but with slower dynamics. Finally, spectral\u0000analysis reveals a blueshift of the THz emission related to the formation of\u0000magnetic domains close to magnetization compensation.","PeriodicalId":501083,"journal":{"name":"arXiv - PHYS - Applied Physics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142177804","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Sebastian Lukas, Nico Rademacher, Sofía Cruces, Michael Gross, Eva Desgué, Stefan Heiserer, Nikolas Dominik, Maximilian Prechtl, Oliver Hartwig, Cormac Ó Coileáin, Tanja Stimpel Lindner, Pierre Legagneux, Arto Rantala, Juha Matti Saari, Miika Soikkeli, Georg S. Duesberg, Max C. Lemme
Membrane-based sensors are an important market for microelectromechanical systems (MEMS). Two-dimensional (2D) materials, with their low mass, are excellent candidates for suspended membranes to provide high sensitivity, small footprint sensors. The present work demonstrates pressure sensors employing large-scale-synthesized 2D platinum diselenide (PtSe${_2}$) films as piezoresistive membranes supported only by a thin polymer layer. We investigate three different synthesis methods with contrasting growth parameters and establish a reliable high yield fabrication process for suspended PtSe${_2}$/PMMA membranes across sealed cavities. The pressure sensors reproducibly display sensitivities above 6 x 10${^4}$ kPa. We show that the sensitivity clearly depends on the membrane diameter and the piezoresistive gauge factor of the PtSe${_2}$ film. Reducing the total device size by decreasing the number of membranes within a device leads to a significant increase in the area-normalized sensitivity. This allows the manufacturing of pressure sensors with high sensitivity but a much smaller device footprint than the current state-of-the-art MEMS technology. We further integrate PtSe${_2}$ pressure sensors with CMOS technology, improving the technological readiness of PtSe${_2}$-based MEMS and NEMS devices.
{"title":"Piezoresistive PtSe$_2$ pressure sensors with reliable high sensitivity and their integration into CMOS ASIC substrates","authors":"Sebastian Lukas, Nico Rademacher, Sofía Cruces, Michael Gross, Eva Desgué, Stefan Heiserer, Nikolas Dominik, Maximilian Prechtl, Oliver Hartwig, Cormac Ó Coileáin, Tanja Stimpel Lindner, Pierre Legagneux, Arto Rantala, Juha Matti Saari, Miika Soikkeli, Georg S. Duesberg, Max C. Lemme","doi":"arxiv-2409.03053","DOIUrl":"https://doi.org/arxiv-2409.03053","url":null,"abstract":"Membrane-based sensors are an important market for microelectromechanical\u0000systems (MEMS). Two-dimensional (2D) materials, with their low mass, are\u0000excellent candidates for suspended membranes to provide high sensitivity, small\u0000footprint sensors. The present work demonstrates pressure sensors employing\u0000large-scale-synthesized 2D platinum diselenide (PtSe${_2}$) films as\u0000piezoresistive membranes supported only by a thin polymer layer. We investigate\u0000three different synthesis methods with contrasting growth parameters and\u0000establish a reliable high yield fabrication process for suspended\u0000PtSe${_2}$/PMMA membranes across sealed cavities. The pressure sensors\u0000reproducibly display sensitivities above 6 x 10${^4}$ kPa. We show that the\u0000sensitivity clearly depends on the membrane diameter and the piezoresistive\u0000gauge factor of the PtSe${_2}$ film. Reducing the total device size by\u0000decreasing the number of membranes within a device leads to a significant\u0000increase in the area-normalized sensitivity. This allows the manufacturing of\u0000pressure sensors with high sensitivity but a much smaller device footprint than\u0000the current state-of-the-art MEMS technology. We further integrate PtSe${_2}$\u0000pressure sensors with CMOS technology, improving the technological readiness of\u0000PtSe${_2}$-based MEMS and NEMS devices.","PeriodicalId":501083,"journal":{"name":"arXiv - PHYS - Applied Physics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142177808","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Fatigue crack growth in ductile materials is primarily driven by the interaction between damaging and shielding mechanisms. In the Paris regime, the predominant mechanism for retardation is plasticity-induced crack closure (PICC). However, some of the mechanisms behind this phenomenon are still unclear. Identifying and separating the three-dimensional aspect from other shielding aspects during experiments is extremely complex. In this paper, we analyze the crack opening kinematics based on local crack opening displacement measurements in both 2D high-resolution digital image correlation data and 3D finite element simulations. The results confirm that the crack opening stress intensity factor Kop differs along the crack path. We present a new method to determine Kop at the crack front allowing us to identify PICC as the predominant shielding mechanism in fatigue crack growth experiments. Furthermore, this work contributes to the discussion on the damage-reducing effect of PICC, since we find that the influence on fatigue damage in the plastic zone remains negligible when the crack is closed and crack surface contact is directed towards the surface.
韧性材料的疲劳裂纹增长主要是由破坏机制和屏蔽机制之间的相互作用所驱动的。在巴黎机制中，主要的延缓机制是塑性诱导的裂纹闭合（PICC）。然而，这一现象背后的一些机制仍不清楚。在实验过程中，识别和区分三维方面与其他屏蔽方面是极其复杂的。在本文中，我们根据二维高分辨率数字图像相关数据和三维有限元模拟中的局部裂纹张开位移测量结果，对裂纹张开运动学进行了分析。结果证实，裂纹开口应力强度因子 Kop 沿裂纹路径不同。我们提出了一种确定裂纹前沿 Kop 的新方法，从而使我们能够确定 PICC 是疲劳裂纹生长实验中的主要屏蔽机制。此外，这项工作还有助于讨论 PICC 的减损效应，因为我们发现当裂纹闭合且裂纹表面接触朝向表面时，PICC 对塑性区疲劳损伤的影响仍然可以忽略不计。
{"title":"Plasticity-induced crack closure identification during fatigue crack growth in AA2024-T3 by using high-resolution digital image correlation","authors":"Florian Paysan, David Melching, Eric Breibarth","doi":"arxiv-2409.02560","DOIUrl":"https://doi.org/arxiv-2409.02560","url":null,"abstract":"Fatigue crack growth in ductile materials is primarily driven by the\u0000interaction between damaging and shielding mechanisms. In the Paris regime, the\u0000predominant mechanism for retardation is plasticity-induced crack closure\u0000(PICC). However, some of the mechanisms behind this phenomenon are still\u0000unclear. Identifying and separating the three-dimensional aspect from other\u0000shielding aspects during experiments is extremely complex. In this paper, we\u0000analyze the crack opening kinematics based on local crack opening displacement\u0000measurements in both 2D high-resolution digital image correlation data and 3D\u0000finite element simulations. The results confirm that the crack opening stress\u0000intensity factor Kop differs along the crack path. We present a new method to\u0000determine Kop at the crack front allowing us to identify PICC as the\u0000predominant shielding mechanism in fatigue crack growth experiments.\u0000Furthermore, this work contributes to the discussion on the damage-reducing\u0000effect of PICC, since we find that the influence on fatigue damage in the\u0000plastic zone remains negligible when the crack is closed and crack surface\u0000contact is directed towards the surface.","PeriodicalId":501083,"journal":{"name":"arXiv - PHYS - Applied Physics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142177807","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Acoustic black holes represent a special class of metastructures allowing efficient absorption based on the slow sound principle. The decrease of the wave speed is associated with the spatial variation of acoustic impedance, while the absorption properties are linked to thermoviscous losses induced by the local resonances of the structure. While most of the developments in the field of sonic black holes are dedicated to one-dimensional structures, the current study is concerned with their two-dimensional counterparts. It is shown that the change of the dimensionality results in the change of noise insulation mechanism, which relies on the opening of band-gaps rather then thermoviscous losses. The formation of band-gaps is associated with the strong coupling between the resonators constituting the considered structures. Numerically and experimentally it is shown than the structure is characterized by broad stop-bands in transmission spectra, while the air flow propagation is still allowed. In particular, a realistic application scenario is considered, in which the acoustic noise and the air flow are generated by a fan embedded into a ventilation duct. The obtained results pave the way towards the development of next-level ventilated metamaterials for efficient noise control.
{"title":"Ventilated noise-insulating metamaterials inspired by sonic black holes","authors":"Farid Bikmukhametov, Lana Glazko, Yaroslav Muravev, Dmitrii Pozdeev, Evgeni Vasiliev, Sergey Krasikov, Mariia Krasikova","doi":"arxiv-2409.02731","DOIUrl":"https://doi.org/arxiv-2409.02731","url":null,"abstract":"Acoustic black holes represent a special class of metastructures allowing\u0000efficient absorption based on the slow sound principle. The decrease of the\u0000wave speed is associated with the spatial variation of acoustic impedance,\u0000while the absorption properties are linked to thermoviscous losses induced by\u0000the local resonances of the structure. While most of the developments in the\u0000field of sonic black holes are dedicated to one-dimensional structures, the\u0000current study is concerned with their two-dimensional counterparts. It is shown\u0000that the change of the dimensionality results in the change of noise insulation\u0000mechanism, which relies on the opening of band-gaps rather then thermoviscous\u0000losses. The formation of band-gaps is associated with the strong coupling\u0000between the resonators constituting the considered structures. Numerically and\u0000experimentally it is shown than the structure is characterized by broad\u0000stop-bands in transmission spectra, while the air flow propagation is still\u0000allowed. In particular, a realistic application scenario is considered, in\u0000which the acoustic noise and the air flow are generated by a fan embedded into\u0000a ventilation duct. The obtained results pave the way towards the development\u0000of next-level ventilated metamaterials for efficient noise control.","PeriodicalId":501083,"journal":{"name":"arXiv - PHYS - Applied Physics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142177806","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Kaito Tachi, Kota Suzuki, Kairi Takimoto, Shunsuke Saruwatari, Kiichi Niitsu, Peter Njogu, Hiroki Wakatsuchi
Electromagnetic (EM) multipath interference is difficult to address with passive approaches due to two physical restrictions - the shared frequency of the initial and interfering signals and their variable incident angles. Thus, to address multipath interference, the spatial impedance must be adjusted in response to the incident angles of multiple signals with the same frequency, which is impossible with classic linear time-invariant (LTI) systems. We present a design concept for metasurface-based spatial filters to overcome LTI behavior and suppress multipath interference signals using a time-varying interlocking mechanism without any active biasing systems. The proposed devices are coupled to the first incoming wave to adjust the spatial impedance and suppress delayed waves in the time domain, which is validated numerically and experimentally. This study opens a new avenue for passive yet time-varying selective EM metasystems, enabling the adjustment of spatially complicated EM waves and fields even at the same frequency.
{"title":"Multipath Signal-Selective Metasurface: Passive Time-Varying Interlocking Mechanism to Vary Spatial Impedance for Signals with the Same Frequency","authors":"Kaito Tachi, Kota Suzuki, Kairi Takimoto, Shunsuke Saruwatari, Kiichi Niitsu, Peter Njogu, Hiroki Wakatsuchi","doi":"arxiv-2409.02484","DOIUrl":"https://doi.org/arxiv-2409.02484","url":null,"abstract":"Electromagnetic (EM) multipath interference is difficult to address with\u0000passive approaches due to two physical restrictions - the shared frequency of\u0000the initial and interfering signals and their variable incident angles. Thus,\u0000to address multipath interference, the spatial impedance must be adjusted in\u0000response to the incident angles of multiple signals with the same frequency,\u0000which is impossible with classic linear time-invariant (LTI) systems. We\u0000present a design concept for metasurface-based spatial filters to overcome LTI\u0000behavior and suppress multipath interference signals using a time-varying\u0000interlocking mechanism without any active biasing systems. The proposed devices\u0000are coupled to the first incoming wave to adjust the spatial impedance and\u0000suppress delayed waves in the time domain, which is validated numerically and\u0000experimentally. This study opens a new avenue for passive yet time-varying\u0000selective EM metasystems, enabling the adjustment of spatially complicated EM\u0000waves and fields even at the same frequency.","PeriodicalId":501083,"journal":{"name":"arXiv - PHYS - Applied Physics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142177809","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Jun Ji, Zichen Xi, Bernadeta R. Srijanto, Ivan I. Kravchenko, Ming Jin, Wenjie Xiong, Linbo Shao
Multiply-accumulation (MAC) is a crucial computing operation in signal processing, numerical simulations, and machine learning. This work presents a scalable, programmable, frequency-domain parallel computing leveraging gigahertz (GHz)-frequency acoustic-wave nonlinearities. By encoding data in the frequency domain, a single nonlinear acoustic-wave device can perform a billion arithmetic operations simultaneously. A single device with a footprint of 0.03 mm$^2$ on lithium niobate (LN) achieves 0.0144 tera floating-point operations per second (TFLOPS), leading to a computing area density of 0.48 TFLOPS/mm$^2$ and a core power efficiency of 0.14 TFLOPS/Watt. As applications, we demonstrate multiplications of two 16-by-16 matrices and convolutional imaging processing of 128-by-128-pixel photos. Our technology could find versatile applications in near-sensor signal processing and edge computing.
{"title":"Frequency-domain Parallel Computing Using Single On-Chip Nonlinear Acoustic-wave Device","authors":"Jun Ji, Zichen Xi, Bernadeta R. Srijanto, Ivan I. Kravchenko, Ming Jin, Wenjie Xiong, Linbo Shao","doi":"arxiv-2409.02689","DOIUrl":"https://doi.org/arxiv-2409.02689","url":null,"abstract":"Multiply-accumulation (MAC) is a crucial computing operation in signal\u0000processing, numerical simulations, and machine learning. This work presents a\u0000scalable, programmable, frequency-domain parallel computing leveraging\u0000gigahertz (GHz)-frequency acoustic-wave nonlinearities. By encoding data in the\u0000frequency domain, a single nonlinear acoustic-wave device can perform a billion\u0000arithmetic operations simultaneously. A single device with a footprint of 0.03\u0000mm$^2$ on lithium niobate (LN) achieves 0.0144 tera floating-point operations\u0000per second (TFLOPS), leading to a computing area density of 0.48 TFLOPS/mm$^2$\u0000and a core power efficiency of 0.14 TFLOPS/Watt. As applications, we\u0000demonstrate multiplications of two 16-by-16 matrices and convolutional imaging\u0000processing of 128-by-128-pixel photos. Our technology could find versatile\u0000applications in near-sensor signal processing and edge computing.","PeriodicalId":501083,"journal":{"name":"arXiv - PHYS - Applied Physics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142177833","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Davi Rodrigues, Eleonora Raimondo, Riccardo Tomasello, Mario Carpentieri, Giovanni Finocchio
The electrically readable complex dynamics of robust and scalable magnetic tunnel junctions (MTJs) offer promising opportunities for advancing neuromorphic computing. In this work, we present an MTJ design with a free layer and two polarizers capable of computing the sigmoidal activation function and its gradient at the device level. This design enables both feedforward and backpropagation computations within a single device, extending neuromorphic computing frameworks previously explored in the literature by introducing the ability to perform backpropagation directly in hardware. Our algorithm implementation reveals two key findings: (i) the small discrepancies between the MTJ-generated curves and the exact software-generated curves have a negligible impact on the performance of the backpropagation algorithm, (ii) the device implementation is highly robust to inter-device variation and noise, and (iii) the proposed method effectively supports transfer learning and knowledge distillation. To demonstrate this, we evaluated the performance of an edge computing network using weights from a software-trained model implemented with our MTJ design. The results show a minimal loss of accuracy of only 0.1% for the Fashion MNIST dataset and 2% for the CIFAR-100 dataset compared to the original software implementation. These results highlight the potential of our MTJ design for compact, hardware-based neural networks in edge computing applications, particularly for transfer learning.
{"title":"A design of magnetic tunnel junctions for the deployment of neuromorphic hardware for edge computing","authors":"Davi Rodrigues, Eleonora Raimondo, Riccardo Tomasello, Mario Carpentieri, Giovanni Finocchio","doi":"arxiv-2409.02528","DOIUrl":"https://doi.org/arxiv-2409.02528","url":null,"abstract":"The electrically readable complex dynamics of robust and scalable magnetic\u0000tunnel junctions (MTJs) offer promising opportunities for advancing\u0000neuromorphic computing. In this work, we present an MTJ design with a free\u0000layer and two polarizers capable of computing the sigmoidal activation function\u0000and its gradient at the device level. This design enables both feedforward and\u0000backpropagation computations within a single device, extending neuromorphic\u0000computing frameworks previously explored in the literature by introducing the\u0000ability to perform backpropagation directly in hardware. Our algorithm\u0000implementation reveals two key findings: (i) the small discrepancies between\u0000the MTJ-generated curves and the exact software-generated curves have a\u0000negligible impact on the performance of the backpropagation algorithm, (ii) the\u0000device implementation is highly robust to inter-device variation and noise, and\u0000(iii) the proposed method effectively supports transfer learning and knowledge\u0000distillation. To demonstrate this, we evaluated the performance of an edge\u0000computing network using weights from a software-trained model implemented with\u0000our MTJ design. The results show a minimal loss of accuracy of only 0.1% for\u0000the Fashion MNIST dataset and 2% for the CIFAR-100 dataset compared to the\u0000original software implementation. These results highlight the potential of our\u0000MTJ design for compact, hardware-based neural networks in edge computing\u0000applications, particularly for transfer learning.","PeriodicalId":501083,"journal":{"name":"arXiv - PHYS - Applied Physics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142177813","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Benjamin M. Goldsberry, Andrew N. Norris, Samuel P. Wallen, Michael R. Haberman
The forced time harmonic response of a spatiotemporally-modulated elastic beam of finite length with light damping is derived using a novel Green's function approach. Closed-form solutions are found that highlight unique mode coupling effects that are induced by spatiotemporal modulation, such as split resonances that are tunable with the modulation parameters. These effects of order unity are caused by spatiotemporal modulation with small amplitude appropriately scaled to the magnitude of the light damping. The scalings identified here between the modulation amplitude, the damping, and the inner range of frequency near the modified resonances, translate over to more complicated and higher dimensional elastic systems.
{"title":"Green's Function Approach to Model Vibrations of Beams with Spatiotemporally Modulated Properties","authors":"Benjamin M. Goldsberry, Andrew N. Norris, Samuel P. Wallen, Michael R. Haberman","doi":"arxiv-2409.02829","DOIUrl":"https://doi.org/arxiv-2409.02829","url":null,"abstract":"The forced time harmonic response of a spatiotemporally-modulated elastic\u0000beam of finite length with light damping is derived using a novel Green's\u0000function approach. Closed-form solutions are found that highlight unique mode\u0000coupling effects that are induced by spatiotemporal modulation, such as split\u0000resonances that are tunable with the modulation parameters. These effects of\u0000order unity are caused by spatiotemporal modulation with small amplitude\u0000appropriately scaled to the magnitude of the light damping. The scalings\u0000identified here between the modulation amplitude, the damping, and the inner\u0000range of frequency near the modified resonances, translate over to more\u0000complicated and higher dimensional elastic systems.","PeriodicalId":501083,"journal":{"name":"arXiv - PHYS - Applied Physics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142177805","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Hot carrier solar cell is proposed where charge carriers are cooled adiabatically in the charge transport layers adjoining the absorber. The device resembles an ideal thermoelectric converter where thermopower, and therefore also carrier entropy, are maintained constant during cooling from the temperature attained in the absorber to the temperature at contacts.
{"title":"Hot carrier solar cells by adiabatic cooling","authors":"Tom Markvart","doi":"arxiv-2409.01716","DOIUrl":"https://doi.org/arxiv-2409.01716","url":null,"abstract":"Hot carrier solar cell is proposed where charge carriers are cooled\u0000adiabatically in the charge transport layers adjoining the absorber. The device\u0000resembles an ideal thermoelectric converter where thermopower, and therefore\u0000also carrier entropy, are maintained constant during cooling from the\u0000temperature attained in the absorber to the temperature at contacts.","PeriodicalId":501083,"journal":{"name":"arXiv - PHYS - Applied Physics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142177812","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：