اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدHardware realization of the multiply and accumulate operation on radio-frequency signals with magnetic tunnel junctions
160
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
Artificial neural networks are a valuable tool for radio-frequency (RF) signal classification in many applications, but digitization of analog signals and the use of general purpose hardware non-optimized for training make the process slow and energetically costly. Recent theoretical work has proposed to use nano-devices called magnetic tunnel junctions, which exhibit intrinsic RF dynamics, to implement in hardware the Multiply and Accumulate (MAC) operation, a key building block of neural networks, directly using analogue RF signals. In this article, we experimentally demonstrate that a magnetic tunnel junction can perform multiplication of RF powers, with tunable positive and negative synaptic weights. Using two magnetic tunnel junctions connected in series we demonstrate the MAC operation and use it for classification of RF signals. These results open the path to embedded systems capable of analyzing RF signals with neural networks directly after the antenna, at low power cost and high speed.
قيم البحث
اقرأ أيضاً
Exploiting the physics of nanoelectronic devices is a major lead for implementing compact, fast, and energy efficient artificial intelligence. In this work, we propose an original road in this direction, where assemblies of spintronic resonators used
as artificial synapses can classify an-alogue radio-frequency signals directly without digitalization. The resonators convert the ra-dio-frequency input signals into direct voltages through the spin-diode effect. In the process, they multiply the input signals by a synaptic weight, which depends on their resonance fre-quency. We demonstrate through physical simulations with parameters extracted from exper-imental devices that frequency-multiplexed assemblies of resonators implement the corner-stone operation of artificial neural networks, the Multiply-And-Accumulate (MAC), directly on microwave inputs. The results show that even with a non-ideal realistic model, the outputs obtained with our architecture remain comparable to that of a traditional MAC operation. Us-ing a conventional machine learning framework augmented with equations describing the physics of spintronic resonators, we train a single layer neural network to classify radio-fre-quency signals encoding 8x8 pixel handwritten digits pictures. The spintronic neural network recognizes the digits with an accuracy of 99.96 %, equivalent to purely software neural net-works. This MAC implementation offers a promising solution for fast, low-power radio-fre-quency classification applications, and a new building block for spintronic deep neural net-works.
Magnetic tunnel junctions operating in the superparamagnetic regime are promising devices in the field of probabilistic computing, which is suitable for applications like high-dimensional optimization or sampling problems. Further, random number gene
ration is of interest in the field of cryptography. For such applications, a devices uncorrelated fluctuation time-scale can determine the effective system speed. It has been theoretically proposed that a magnetic tunnel junction designed to have only easy-plane anisotropy provides fluctuation rates determined by its easy-plane anisotropy field, and can perform on nanosecond or faster time-scale as measured by its magnetoresistances autocorrelation in time. Here we provide experimental evidence of nanosecond scale fluctuations in a circular shaped easy-plane magnetic tunnel junction, consistent with finite-temperature coupled macrospin simulation results and prior theoretical expectations. We further assess the degree of stochasticity of such signal.
Magnetic tunnel junctions (MTJs) are basic building blocks for devices such as magnetic random access memories (MRAMs). The relevance for modern computation of non-volatile high-frequency memories makes ac-transport measurements of MTJs crucial for e
xploring this regime. Here we demonstrate a frequency-mediated effect in which the tunnel magnetoimpedance reverses its sign in a classical Co/Al{_2}O{_3}/NiFe MTJ, whereas we only observe a gradual decrease of tunnel magnetophase. Such effects are explained by the capacitive coupling of a parallel resistor and capacitor in the equivalent circuit model of the MTJ. Furthermore, we report a positive tunnel magnetocapacitance effect, suggesting the presence of a spin-capacitance at the two ferromagnet/tunnel-barrier interfaces. Our results are important for understanding spin transport phenomena at the high frequency regime, in which the spin-polarized charge accumulation at the two interfaces plays a crucial role.
Complex oxide systems have attracted considerable attention because of their fascinating properties, including the magnetic ordering at the conducting interface between two band insulators, such as LaAlO3 (LAO) and SrTiO3 (STO). However, the manipula
tion of the spin degree of freedom at the LAO/STO heterointerface has remained elusive. Here, we have fabricated hybrid magnetic tunnel junctions consisting of Co and LAO/STO ferromagnets with the insertion of a Ti layer in between, which clearly exhibit magnetic switching and the tunnelling magnetoresistance (TMR) effect below 10 K. The magnitude and the of the TMR are strongly dependent on the direction of the rotational magnetic field parallel to the LAO/STO plane, which is attributed to a strong Rashba-type spin orbit coupling in the LAO/STO heterostructure. Our study provides a further support for the existence of the macroscopic ferromagnetism at LAO/STO heterointerfaces and opens a novel route to realize interfacial spintronics devices.
The transport properties of magnetic tunnel junctions (MTJs) are very sensitive to interface modifications. In this work we investigate both experimentally and theoretically the effect of asymmetric barrier modifications on the bias dependence of tun
neling magnetoresistance (TMR) in single crystal Fe/MgO-based MTJs with (i) one crystalline and one rough interface and (ii) with a monolayer of O deposited at the crystalline interface. In both cases we observe an asymmetric bias dependence of TMR and a reversal of its sign at large bias. We propose a general model to explain the bias dependence in these and similar systems reported earlier. The model predicts the existence of two distinct TMR regimes: (i) tunneling regime when the interface is modified with layers of a different insulator and (ii) resonant regime when thin metallic layers are inserted at the interface. We demonstrate that in the tunneling regime negative TMR is due to the high voltage which overcomes the exchange splitting in the electrodes, while the asymmetric bias dependence of TMR is due to the interface transmission probabilities. In the resonant regime inversion of TMR could happen at zero voltage depending on the alignment of the resonance levels with the Fermi surfaces of the electrodes. Moreover, the model predicts a regime in which TMR has different sign at positive and negative bias suggesting possibilities of combining memory with logic functions.
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
