arXiv subjects

Zhihong Chen

Publications and source records attributed to Zhihong Chen.

At least 73 records · Page 4Linked to original sources

arXiv2020

Spin orbit torque controlled stochastic oscillators with synchronization and frequency tunability

Stochastic oscillators based on emerging nanodevices are attractive because of their ultra-low power requirement and ability to exhibit stochastic resonance, a phenomenon where synchronization to weak input signals is enabled due to ambient noise. In this work, a low barrier nanomagnet based stochastic oscillator is demonstrated, whose output jumps spontaneously between two states by harnessing the ambient thermal noise, requiring no additional power. Utilizing spin orbit torque in a three terminal device configuration, phase synchronization of these oscillators to weak periodic drives of particular frequencies is demonstrated. Experiments are performed to show the tunability of this synchronization frequency by controlling an electrical feedback parameter. The current required for synchronization is more than 8 times smaller than that required for deterministic switching of similar nanomagnetic devices. A model based on Kramers' transition rate in a symmetric double well potential is adopted and dynamical simulations are performed to explain the experimental results.

cond-mat.mes-hall↗

arXiv2020

Attention-Guided Discriminative Region Localization and Label Distribution Learning for Bone Age Assessment

Bone age assessment (BAA) is clinically important as it can be used to diagnose endocrine and metabolic disorders during child development. Existing deep learning based methods for classifying bone age use the global image as input, or exploit local information by annotating extra bounding boxes or key points. However, training with the global image underutilizes discriminative local information, while providing extra annotations is expensive and subjective. In this paper, we propose an attention-guided approach to automatically localize the discriminative regions for BAA without any extra annotations. Specifically, we first train a classification model to learn the attention maps of the discriminative regions, finding the hand region, the most discriminative region (the carpal bones), and the next most discriminative region (the metacarpal bones). Guided by those attention maps, we then crop the informative local regions from the original image and aggregate different regions for BAA. Instead of taking BAA as a general regression task, which is suboptimal due to the label ambiguity problem in the age label space, we propose using joint age distribution learning and expectation regression, which makes use of the ordinal relationship among hand images with different individual ages and leads to more robust age estimation. Extensive experiments are conducted on the RSNA pediatric bone age data set. Using no training annotations, our method achieves competitive results compared with existing state-of-the-art semi-automatic deep learning-based methods that require manual annotation. Code is available at https: //github.com/chenchao666/Bone-Age-Assessment.

cs.CV↗

arXiv2020

ESAM: Discriminative Domain Adaptation with Non-Displayed Items to Improve Long-Tail Performance

Most of ranking models are trained only with displayed items (most are hot items), but they are utilized to retrieve items in the entire space which consists of both displayed and non-displayed items (most are long-tail items). Due to the sample selection bias, the long-tail items lack sufficient records to learn good feature representations, i.e. data sparsity and cold start problems. The resultant distribution discrepancy between displayed and non-displayed items would cause poor long-tail performance. To this end, we propose an entire space adaptation model (ESAM) to address this problem from the perspective of domain adaptation (DA). ESAM regards displayed and non-displayed items as source and target domains respectively. Specifically, we design the attribute correlation alignment that considers the correlation between high-level attributes of the item to achieve distribution alignment. Furthermore, we introduce two effective regularization strategies, i.e. \textit{center-wise clustering} and \textit{self-training} to improve DA process. Without requiring any auxiliary information and auxiliary domains, ESAM transfers the knowledge from displayed items to non-displayed items for alleviating the distribution inconsistency. Experiments on two public datasets and a large-scale industrial dataset collected from Taobao demonstrate that ESAM achieves state-of-the-art performance, especially in the long-tail space. Besides, we deploy ESAM to the Taobao search engine, leading to significant improvement on online performance. The code is available at \url{https://github.com/A-bone1/ESAM.git}

cs.IR↗

arXiv2020

Hardware implementation of Bayesian network building blocks with stochastic spintronic devices

Bayesian networks are powerful statistical models to understand causal relationships in real-world probabilistic problems such as diagnosis, forecasting, computer vision, etc. For systems that involve complex causal dependencies among many variables, the complexity of the associated Bayesian networks become computationally intractable. As a result, direct hardware implementation of these networks is one promising approach to reducing power consumption and execution time. However, the few hardware implementations of Bayesian networks presented in literature rely on deterministic CMOS devices that are not efficient in representing the inherently stochastic variables in a Bayesian network. This work presents an experimental demonstration of a Bayesian network building block implemented with naturally stochastic spintronic devices. These devices are based on nanomagnets with perpendicular magnetic anisotropy, initialized to their hard axes by the spin orbit torque from a heavy metal under-layer utilizing the giant spin Hall effect, enabling stochastic behavior. We construct an electrically interconnected network of two stochastic devices and manipulate the correlations between their states by changing connection weights and biases. By mapping given conditional probability tables to the circuit hardware, we demonstrate that any two node Bayesian networks can be implemented by our stochastic network. We then present the stochastic simulation of an example case of a four node Bayesian network using our proposed device, with parameters taken from the experiment. We view this work as a first step towards the large scale hardware implementation of Bayesian networks.

cond-mat.mes-hall↗

arXiv2020

Electrically-Tunable Stochasticity for Spin-based Neuromorphic Circuits: Self-Adjusting to Variation

Energy-efficient methods are addressed for leveraging low energy barrier nanomagnetic devices within neuromorphic architectures. Using a Magnetoresistive Random Access Memory (MRAM) probabilistic device (p-bit) as the basis of neuronal structures in Deep Belief Networks (DBNs), the impact of reducing the Magnetic Tunnel Junction's (MTJ's) energy barrier is assessed and optimized for the resulting stochasticity present in the learning system. This can mitigate the process variation sensitivity of stochastic DBNs which encounter a sharp drop-off when energy barriers exceed near-zero kT. As evaluated for the MNIST dataset for energy barriers at near-zero kT to 2.0 kT in increments of 0.5 kT, it is shown that the stability factor changes by 5 orders of magnitude. The self-compensating circuit developed herein provides a compact, and low complexity approach to mitigating process variation impacts towards practical implementation and fabrication.

cs.ET↗

arXiv2020

Selective Transfer with Reinforced Transfer Network for Partial Domain Adaptation

One crucial aspect of partial domain adaptation (PDA) is how to select the relevant source samples in the shared classes for knowledge transfer. Previous PDA methods tackle this problem by re-weighting the source samples based on their high-level information (deep features). However, since the domain shift between source and target domains, only using the deep features for sample selection is defective. We argue that it is more reasonable to additionally exploit the pixel-level information for PDA problem, as the appearance difference between outlier source classes and target classes is significantly large. In this paper, we propose a reinforced transfer network (RTNet), which utilizes both high-level and pixel-level information for PDA problem. Our RTNet is composed of a reinforced data selector (RDS) based on reinforcement learning (RL), which filters out the outlier source samples, and a domain adaptation model which minimizes the domain discrepancy in the shared label space. Specifically, in the RDS, we design a novel reward based on the reconstruct errors of selected source samples on the target generator, which introduces the pixel-level information to guide the learning of RDS. Besides, we develope a state containing high-level information, which used by the RDS for sample selection. The proposed RDS is a general module, which can be easily integrated into existing DA models to make them fit the PDA situation. Extensive experiments indicate that RTNet can achieve state-of-the-art performance for PDA tasks on several benchmark datasets.

cs.LG↗

arXiv2020

Ro-SOS: Metric Expression Network (MEnet) for Robust Salient Object Segmentation

Although deep CNNs have brought significant improvement to image saliency detection, most CNN based models are sensitive to distortion such as compression and noise. In this paper, we propose an end-to-end generic salient object segmentation model called Metric Expression Network (MEnet) to deal with saliency detection with the tolerance of distortion. Within MEnet, a new topological metric space is constructed, whose implicit metric is determined by the deep network. As a result, we manage to group all the pixels in the observed image semantically within this latent space into two regions: a salient region and a non-salient region. With this architecture, all feature extractions are carried out at the pixel level, enabling fine granularity of output boundaries of the salient objects. What's more, we try to give a general analysis for the noise robustness of the network in the sense of Lipschitz and Jacobian literature. Experiments demonstrate that robust salient maps facilitating object segmentation can be generated by the proposed metric. Tests on several public benchmarks show that MEnet has achieved desirable performance. Furthermore, by direct computation and measuring the robustness, the proposed method outperforms previous CNN-based methods on distorted inputs.

cs.CV↗

arXiv2019

HoMM: Higher-order Moment Matching for Unsupervised Domain Adaptation

Minimizing the discrepancy of feature distributions between different domains is one of the most promising directions in unsupervised domain adaptation. From the perspective of distribution matching, most existing discrepancy-based methods are designed to match the second-order or lower statistics, which however, have limited expression of statistical characteristic for non-Gaussian distributions. In this work, we explore the benefits of using higher-order statistics (mainly refer to third-order and fourth-order statistics) for domain matching. We propose a Higher-order Moment Matching (HoMM) method, and further extend the HoMM into reproducing kernel Hilbert spaces (RKHS). In particular, our proposed HoMM can perform arbitrary-order moment tensor matching, we show that the first-order HoMM is equivalent to Maximum Mean Discrepancy (MMD) and the second-order HoMM is equivalent to Correlation Alignment (CORAL). Moreover, the third-order and the fourth-order moment tensor matching are expected to perform comprehensive domain alignment as higher-order statistics can approximate more complex, non-Gaussian distributions. Besides, we also exploit the pseudo-labeled target samples to learn discriminative representations in the target domain, which further improves the transfer performance. Extensive experiments are conducted, showing that our proposed HoMM consistently outperforms the existing moment matching methods by a large margin. Codes are available at \url{https://github.com/chenchao666/HoMM-Master}

cs.CV↗

arXiv2019

Valley-Coupled-Spintronic Non-Volatile Memories with Compute-In-Memory Support

In this work, we propose valley-coupled spin-hall memories (VSH-MRAMs) based on monolayer WSe2. The key features of the proposed memories are (a) the ability to switch magnets with perpendicular magnetic anisotropy (PMA) via VSH effect and (b) an integrated gate that can modulate the charge/spin current (IC/IS) flow. The former attribute results in high energy efficiency (compared to the Giant-Spin Hall (GSH) effect-based devices with in-plane magnetic anisotropy (IMA) magnets). The latter feature leads to a compact access transistor-less memory array design. We experimentally measure the gate controllability of the current as well as the nonlocal resistance associated with VSH effect. Based on the measured data, we develop a simulation framework (using physical equations) to propose and analyze single-ended and differential VSH effect based magnetic memories (VSH-MRAM and DVSH-MRAM, respectively). At the array level, the proposed VSH/DVSH-MRAMs achieve 50%/ 11% lower write time, 59%/ 67% lower write energy and 35%/ 41% lower read energy at iso-sense margin, compared to single ended/differential (GSH/DGSH)-MRAMs. System level evaluation in the context of general purpose processor and intermittently-powered system shows up to 3.14X and 1.98X better energy efficiency for the proposed (D)VSH-MRAMs over (D)GSH-MRAMs respectively. Further, the differential sensing of the proposed DVSH-MRAM leads to natural and simultaneous in-memory computation of bit-wise AND and NOR logic functions. Using this feature, we design a computation-in-memory (CiM) architecture that performs Boolean logic and addition (ADD) with a single array access. System analysis performed by integrating our DVSH-MRAM: CiM in the Nios II processor across various application benchmarks shows up to 2.66X total energy savings, compared to DGSH-MRAM: CiM.

cs.ET↗

arXiv2019

Atomically Controlled Tunable Doping in High Performance WSe2 Devices

Two-dimensional transitional metal dichalcogenide (TMD) field-effect transistors (FETs) are promising candidates for future electronic applications, owing to their excellent transport properties and potential for ultimate device scaling. However, it is widely acknowledged that substantial contact resistance associated with the contact-TMD interface has impeded device performance to a large extent. It has been discovered that O2 plasma treatment can convert WSe2 into WO3-x and substantially improve contact resistances of p-type WSe2 devices by strong doping induced thinner depletion width. In this paper, we carefully study the temperature dependence of this conversion, demonstrating an oxidation process with a precise monolayer control at room temperature and multilayer conversion at elevated temperatures. Furthermore, the lateral oxidation of WSe2 under the contact revealed by HR-STEM leads to potential unpinning of the metal Fermi level and Schottky barrier lowering, resulting in lower contact resistances. The p-doping effect is attributed to the high electron affinity of the formed WO3-x layer on top of the remaining WSe2 channel, and the doping level is found to be dependent on the WO3-x thickness that is controlled by the temperature. Comprehensive materials and electrical characterizations are presented, with a low contact resistance of ~528 ohm-um and record high on-state current of 320 uA/um at -1V bias being reported.

physics.app-ph↗

arXiv2019

Correlated fluctuations in spin orbit torque-coupled perpendicular nanomagnets

Low barrier nanomagnets have attracted a lot of research interest for their use as sources of high quality true random number generation. More recently, low barrier nanomagnets with tunable output have been shown to be a natural hardware platform for unconventional computing paradigms such as probabilistic spin logic. Efficient generation and tunability of high quality random bits is critical for these novel applications. However, current spintronic random number generators are based on superparamagnetic tunnel junctions (SMTJs) with tunability obtained through spin transfer torque (STT), which unavoidably leads to challenges in designing concatenated networks using these two terminal devices. The more recent development of utilizing spin orbit torque (SOT) allows for a three terminal device design, but can only tune in-plane magnetization freely, which is not very energy efficient due to the needs of overcoming a large demagnetization field. In this work, we experimentally demonstrate for the first time, a stochastic device with perpendicular magnetic anisotropy (PMA) that is completely tunable by SOT without the aid of any external magnetic field. Our measurements lead us to hypothesize that a tilted anisotropy might be responsible for the observed tunability. We carry out stochastic Landau-Lifshitz-Gilbert (sLLG) simulations to confirm our experimental observation. Finally, we build an electrically coupled network of two such stochastic nanomagnet based devices and demonstrate that finite correlation or anti-correlation can be established between their output fluctuations by a weak interconnection, despite having a large difference in their natural fluctuation time scale. Simulations based on a newly developed dynamical model for autonomous circuits composed of low barrier nanomagnets show close agreement with the experimental results.

physics.app-ph↗

arXiv2019

Experimental observation of coupled valley and spin Hall effect in p-doped WSe2 devices

Giant spin Hall effect (GSHE) has been observed in heavy metal materials such as Ta, Pt, and W, where spins are polarized in the surface plane and perpendicular to the charge current direction. Spins generated in these materials have successfully switched magnets with in-plane magnetic anisotropy (IMA) and perpendicular magnetic anisotropy (PMA) through spin orbit torque (SOT) mechanism. It is generally accepted that PMA magnets are preferred over IMA magnets in data storage applications owing to their large thermal stability even at ultra scaled dimensions. However, SOT switching of PMA magnets by conventional GSHE materials requires either a small external magnetic field, a local dipolar field, or introducing tilted anisotropy to break the symmetry with respect to the magnetization. To deterministically switch a PMA without any additional assistance, nonconventional GSHE materials that can generate spins with polarization perpendicular to the surfaces are needed. Several monolayer transition metal dichalcogenides (TMDs) have been predicted to generate such out of plane spins due to their 2D nature and unique band structures. Interestingly, opposite spins are locked to their respective sub-band in each K valley of the TMD valence band with substantially large energy splitting, which enables polarized spins to be accessible through electrical gating and spatially separated by electric field through the valley Hall effect (VHE). Therefore, spatial separation and accumulation of spins in these 2D TMDs are uniquely referred to as coupled valley and spin Hall effect. Here, we report an experiment of electrical generation of spin current with out of plane polarization in monolayer WSe2 and detection of spin signals through a nonlocal spin valve structure built on a lateral graphene spin diffusion channel that partially overlaps with WSe2.

cond-mat.mes-hall↗

arXiv2019

Towards Self-similarity Consistency and Feature Discrimination for Unsupervised Domain Adaptation

Recent advances in unsupervised domain adaptation mainly focus on learning shared representations by global distribution alignment without considering class information across domains. The neglect of class information, however, may lead to partial alignment (or even misalignment) and poor generalization performance. For comprehensive alignment, we argue that the similarities across different features in the source domain should be consistent with that of in the target domain. Based on this assumption, we propose a new domain discrepancy metric, i.e., Self-similarity Consistency (SSC), to enforce the feature structure being consistent across domains. The renowned correlation alignment (CORAL) is proven to be a special case, and a sub-optimal measure of our proposed SSC. Furthermore, we also propose to mitigate the side effect of the partial alignment and misalignment by incorporating the discriminative information of the deep representations. Specifically, an embarrassingly simple and effective feature norm constraint is exploited to enlarge the discrepancy of inter-class samples. It relieves the requirements of strict alignment when performing adaptation, therefore improving the adaptation performance significantly. Extensive experiments on visual domain adaptation tasks demonstrate the effectiveness of our proposed SSC metric and feature discrimination approach.

cs.CV↗

arXiv2019

Enhancing Interconnect Reliability and Performance by Converting Tantalum to 2D Layered Tantalum Sulfide at Low Temperature

The interconnect half-pitch size will reach ~20 nm in the coming sub-5 nm technology node. Meanwhile, the TaN/Ta (barrier/liner) bilayer stack has to be > 4 nm to ensure acceptable liner and diffusion barrier properties. Since TaN/Ta occupy a significant portion of the interconnect cross-section and they are much more resistive than Cu, the effective conductance of an ultra-scaled interconnect will be compromised by the thick bilayer. Therefore, two dimensional (2D) layered materials have been explored as diffusion barrier alternatives. However, many of the proposed 2D barriers are prepared at too high temperatures to be compatible with the back-end-of-line (BEOL) technology. In addition, as important as the diffusion barrier properties, the liner properties of 2D materials must be evaluated, which has not yet been pursued. Here, a 2D layered tantalum sulfide (TaSx) with ~1.5 nm thickness is developed to replace the conventional TaN/Ta bilayer. The TaSx ultra-thin film is industry-friendly, BEOL-compatible, and can be directly prepared on dielectrics. Our results show superior barrier/liner properties of TaSx compared to the TaN/Ta bilayer. This single-stack material, serving as both a liner and a barrier, will enable continued scaling of interconnects beyond 5 nm node.

cond-mat.mtrl-sci↗

arXiv2018

Joint Domain Alignment and Discriminative Feature Learning for Unsupervised Deep Domain Adaptation

Recently, considerable effort has been devoted to deep domain adaptation in computer vision and machine learning communities. However, most of existing work only concentrates on learning shared feature representation by minimizing the distribution discrepancy across different domains. Due to the fact that all the domain alignment approaches can only reduce, but not remove the domain shift. Target domain samples distributed near the edge of the clusters, or far from their corresponding class centers are easily to be misclassified by the hyperplane learned from the source domain. To alleviate this issue, we propose to joint domain alignment and discriminative feature learning, which could benefit both domain alignment and final classification. Specifically, an instance-based discriminative feature learning method and a center-based discriminative feature learning method are proposed, both of which guarantee the domain invariant features with better intra-class compactness and inter-class separability. Extensive experiments show that learning the discriminative features in the shared feature space can significantly boost the performance of deep domain adaptation methods.

cs.LG↗

arXiv2018

Studies of two-dimensional MoS2 on enhancing the electrical performance of ultrathin copper films

Copper nanowires are widely used as on-chip interconnects due to superior conductivity. However, with aggressive Cu interconnect scaling, the diffusive surface scattering of electrons drastically increases the electrical resistivity. In this work, we studied the electrical performance of Cu thin films on different materials. By comparing the thickness dependence of Cu films resistivity on MoS2 and SiO2, we demonstrated that two-dimensional MoS2 can be used to enhance the electrical performance of ultrathin Cu films due to a partial specular surface scattering. By fitting the experimental data with the theoretical Fuchs Sondheimer model, we obtained the specularity parameter at the Cu MoS2 interface in the temperature range 2K to 300K. Furthermore, first principle calculations based on the density functional theory indicates that there are more localized states at the Cu amorphous SiO2 interface than the Cu MoS2 interface which is responsible for the higher resistivity in the Cu SiO2 heterostructure due to more severe electron scattering. Our results suggest that Cu MoS2 hybrid is a promising candidate structure for the future generations of CMOS interconnects.

physics.app-ph↗

arXiv2018

Direct Observation of Valley Coupled Topological Current in MoS$_2$

The valley degree of freedom of electrons in two-dimensional transition metal dichalcogenides has been extensively studied by theory, optical and optoelectronic experiments. However, generation and detection of pure valley current without relying on optical selection have not yet been demonstrated in these materials. Here, we report that valley current can be electrically induced and detected through the valley Hall effect and inverse valley Hall effect, respectively, in monolayer molybdenum disulfide. Specifically, long-range valley transport is observed over half a micron distance at room temperature. Our findings will enable a new generation of electronic devices utilizing the valley degree of freedom, which can be used for future novel valleytronic applications.

cond-mat.mes-hall↗

arXiv2017

Two-Dimensional h-BN and MoS2 as Diffusion Barriers for Ultra-Scaled Copper Interconnects

Copper interconnects in modern integrated circuits require ultra-thin barriers to prevent intermixing of Cu with surrounding dielectric materials. Conventional barriers rely on metals like TaN, however their finite thickness reduces the cross-sectional area and significantly increases the resistivity of nanoscale interconnects. In this study, a new class of two-dimensional (2D) Cu diffusion barriers, hexagonal boron nitride (h-BN) and molybdenum disulfide (MoS2), is demonstrated for the first time. Using time-dependent dielectric breakdown measurements and scanning transmission electron microscopy coupled with energy dispersive X-ray spectroscopy and electron energy loss spectroscopy, these 2D materials are shown to be promising barrier solutions for ultra-scaled interconnect technology. The predicted lifetime of devices with directly deposited 2D barriers can achieve three orders of magnitude improvement compared to control devices without barriers.

cond-mat.mtrl-sci↗