arXiv subjects

Lin Gu

Publications and source records attributed to Lin Gu.

At least 145 records · Page 8Linked to original sources

arXiv2021

Linear magnetization dependence and large intrinsic anomalous Hall effect in Fe78Si9B13 metallic glasses

The origin of anomalous Hall effect (AHE) in ferromagnetic metallic glasses (MGs) is not yet understood completely. Here, the AHE is explored in Fe78Si9B13 MGs. We find the behavior of resistivity at low temperature seems to be more likely due to structure effect rather than Kondo-type effect. More importantly, we firstly find the primitive experiment anomalous Hall conductivity (σAH) without separation of extrinsic contribution has a linear magnetization (Mz) dependence when temperature is changing, which is another feature of intrinsic mechanism and indicates intrinsic contribution is dominated. Furthermore, the σAH normalized by Mz is independent of longitudinal conductivity (σxx), which shows the characteristic of dissipationless intrinsic mechanism. We suggest the intrinsic contribution can be understood from the density of Berry curvature integrated over occupied energies proposed for aperiodic materials recently, and the linear magnetization dependence can be understood qualitatively from the fluctuations of spin orientation and the proportional relationship between Berry curvature and magnetization. Moreover, based on the recent theory report of topological amorphous metals, we make a prediction that the large intrinsic σAH (616 S/cm) in Fe78Si9B13 MGs implies some topological properties of MGs waiting for further discovery.

cond-mat.mtrl-sci↗

arXiv2021

Presence of $s$-wave pairing in Josephson junctions made of twisted ultrathin Bi$_2$Sr$_2$CaCu$_2$O$_{8+x}$ flakes

Since the discovery of high temperature superconductivity in cuprates, Josephson junction based phase-sensitive experiments are believed and used to provide the most convincing evidence for determining the pairing symmetry. Regardless of different junction materials and geometries used, quantum tunneling involved in these experiments is essentially a nanoscale process, and thus, actual experimental results are extremely sensitive to atomic details of the junction structures. The situation has led to controversial results as to the nature of the pairing symmetry of cuprates: while in-plane junction experiments generally support $d$-wave pairing symmetry, those based on out-of-plane ($c$-axis) Josephson junctions between two rotated cuprate blocks favor $s$-wave pairing. In this work, we revisit the $c$-axis experiment by fabricating Josephson junctions with atomic-level control in their interface structure. We fabricate over 90 junctions of ultrathin Bi$_2$Sr$_2$CaCu$_2$O$_{8+x}$ (BSCCO) flakes by state-of-the-art exfoliation technique and obtain atomically flat junction interfaces in the whole junction regions as characterized by high resolution transmission electron microscopy (TEM). Notably, the resultant uniform junctions at various twist angles all exhibit a single tunneling branch behavior, suggesting that only the first half a unit cell on both sides of the twisted flakes is involved in Josephson tunneling process. With such well-defined geometry/structure and the characteristic single tunneling branch, we repeatedly observe Josephson tunneling at a nominal twist angle of 45 degrees, which is against the expectation from a purely $d$-wave pairing scenario. Our results strongly favor the scenario of a persistent $s$-wave order parameter in the junction.

cond-mat.supr-con↗

arXiv2021

Explainable Diabetic Retinopathy Detection and Retinal Image Generation

Though deep learning has shown successful performance in classifying the label and severity stage of certain diseases, most of them give few explanations on how to make predictions. Inspired by Koch's Postulates, the foundation in evidence-based medicine (EBM) to identify the pathogen, we propose to exploit the interpretability of deep learning application in medical diagnosis. By determining and isolating the neuron activation patterns on which diabetic retinopathy (DR) detector relies to make decisions, we demonstrate the direct relation between the isolated neuron activation and lesions for a pathological explanation. To be specific, we first define novel pathological descriptors using activated neurons of the DR detector to encode both spatial and appearance information of lesions. Then, to visualize the symptom encoded in the descriptor, we propose Patho-GAN, a new network to synthesize medically plausible retinal images. By manipulating these descriptors, we could even arbitrarily control the position, quantity, and categories of generated lesions. We also show that our synthesized images carry the symptoms directly related to diabetic retinopathy diagnosis. Our generated images are both qualitatively and quantitatively superior to the ones by previous methods. Besides, compared to existing methods that take hours to generate an image, our second level speed endows the potential to be an effective solution for data augmentation.

eess.IV↗

arXiv2021

A magnetic Weyl semimetallic phase in thin films of Eu$_2$Ir$_2$O$_7$

The interplay between electronic interactions and strong spin-orbit coupling is expected to create a plethora of fascinating correlated topological states of quantum matter. Of particular interest are magnetic Weyl semimetals originally proposed in the pyrochlore iridates, which are only expected to reveal their topological nature in thin film form. To date, however, direct experimental demonstrations of these exotic phases remain elusive, due to the lack of usable single crystals and the insufficient quality of available films. Here, we report on the discovery of the long-sought magnetic Weyl semi-metallic phase in (111)-oriented Eu$_2$Ir$_2$O$_7$ high-quality epitaxial thin films. The topological magnetic state shows an intrinsic anomalous Hall effect with colossal coercivity but vanishing net magnetization, which emerges below the onset of a peculiar magnetic phase with all-in-all-out antiferromagnetic ordering. The observed anomalous Hall conductivity arises from the non-zero Berry curvature emanated by Weyl node pairs near the Fermi level that act as sources and sinks of Berry flux, activated by broken cubic crystal symmetry at the top and bottom terminations of the thin film.

cond-mat.str-el↗

arXiv2021

Estimating and Improving Fairness with Adversarial Learning

Fairness and accountability are two essential pillars for trustworthy Artificial Intelligence (AI) in healthcare. However, the existing AI model may be biased in its decision marking. To tackle this issue, we propose an adversarial multi-task training strategy to simultaneously mitigate and detect bias in the deep learning-based medical image analysis system. Specifically, we propose to add a discrimination module against bias and a critical module that predicts unfairness within the base classification model. We further impose an orthogonality regularization to force the two modules to be independent during training. Hence, we can keep these deep learning tasks distinct from one another, and avoid collapsing them into a singular point on the manifold. Through this adversarial training method, the data from the underprivileged group, which is vulnerable to bias because of attributes such as sex and skin tone, are transferred into a domain that is neutral relative to these attributes. Furthermore, the critical module can predict fairness scores for the data with unknown sensitive attributes. We evaluate our framework on a large-scale public-available skin lesion dataset under various fairness evaluation metrics. The experiments demonstrate the effectiveness of our proposed method for estimating and improving fairness in the deep learning-based medical image analysis system.

cs.CV↗

arXiv2021

One-step epitaxy of high-mobility La-doped BaSnO3 films by high-pressure magnetron sputtering

As a unique perovskite transparent oxide semiconductor, high-mobility La-doped BaSnO3 films have been successfully synthesized by molecular beam epitaxy and pulsed laser deposition. However, it remains a big challenge for magnetron sputtering, a widely applied technique suitable for large-scale fabrication, to grow high-mobility La-doped BaSnO3 films. Here, we developed a method to synthesize high-mobility epitaxial La-doped BaSnO3 films (mobility up to 121 cm2V-1s-1 at the carrier density ~ 4.0 x 10^20 cm-3 at room temperature) directly on SrTiO3 single crystal substrates using high-pressure magnetron sputtering. The structural and electrical properties of the La-doped BaSnO3 films were characterized by combined high-resolution X-ray diffraction, X-ray photoemission spectroscopy, and temperature-dependent electrical transport measurements. The room temperature electron mobility of La-doped BaSnO3 films in this work is 2 to 4 times higher than the reported values of the films grown by magnetron sputtering. Moreover, in the high carrier density range (n > 3 x 10^20 cm-3), the electron mobility value of 121 cm2V-1s-1 in our work is among the highest values for all reported doped BaSnO3 films. It is revealed that high argon pressure during sputtering plays a vital role in stabilizing the fully relaxed films and inducing oxygen vacancies, which benefit the high mobility at room temperature. Our work provides an easy and economical way to massively synthesize high-mobility transparent conducting films for transparent electronics.

cond-mat.mtrl-sci↗

arXiv2021

In Operando magnetometry study on the charge storage mechanism of SnCo alloy lithium ion batteries

In view of the long-standing controversy over the reversibility of transition metals in Sn-based alloys as anode for Li-ion batteries, an in situ real-time magnetic monitoring method was used to investigate the evolution of Sn-Co intermetallic during the electrochemical cycling. Sn-Co alloy film anodes with different compositions were prepared via magnetron sputtering without using binders and conductive additives. The magnetic responses showed that the Co particles liberated by Li insertion recombine fully with Sn during the delithiation to reform Sn-Co intermetallic into stannum richer phases Sn7Co3. However, as the Co content increases, it can only recombine partially with Sn into cobalt richer phases Sn3Co7. The unconverted Co particles may form a dense barrier layer and prevent the full reaction of Li with all the Sn in the anode, leading to lower capacities. These critical results shed light on understanding the reaction mechanism of transition metals, and provide valuable insights toward the design of high-performance Sn alloy based anodes.

cond-mat.mtrl-sci↗

arXiv2021

Emergent Magnetic Phenomenon with Unconventional Structure in Epitaxial Manganate Thin Films

A variety of emergent phenomena has been enabled by interface engineering in the complex oxides heterostructures. While extensive attention has been attracted to LaMnO3 (LMO) thin films for observing the control of functionalities at its interface with substrate, the nature of the magnetic phases in the thin film is, however,controversial. Here, it is reported that the ferromagnetism in 2 and 5 unit cells thick LMO films epitaxially deposited (001)-SrTiO3 substrates ferromagnetic/ferromagnetic coupling in 8 and 10-unit-cell ones, and a striking ferromagnetic/antiferromagnetic pinning effect with apparent positive exchange bias in 15 and 20-unit-cell ones are observed. This novel phenomenon in both 15 and 20-unit-cell films indicates a coexistence of three magnetic orderings in a single LMO film.The high-resolution scanning transmission electron microscopy suggests a P21/n to Pbnm symmetry transition from interface to surface, with the spatial stratification of MnO6 octahedral morphology, corresponding to different magnetic orderings. These results should shed some new lights on manipulating the functionality of oxides by interface engineering.

cond-mat.str-el↗

arXiv2021

Goal-Oriented Gaze Estimation for Zero-Shot Learning

Zero-shot learning (ZSL) aims to recognize novel classes by transferring semantic knowledge from seen classes to unseen classes. Since semantic knowledge is built on attributes shared between different classes, which are highly local, strong prior for localization of object attribute is beneficial for visual-semantic embedding. Interestingly, when recognizing unseen images, human would also automatically gaze at regions with certain semantic clue. Therefore, we introduce a novel goal-oriented gaze estimation module (GEM) to improve the discriminative attribute localization based on the class-level attributes for ZSL. We aim to predict the actual human gaze location to get the visual attention regions for recognizing a novel object guided by attribute description. Specifically, the task-dependent attention is learned with the goal-oriented GEM, and the global image features are simultaneously optimized with the regression of local attribute features. Experiments on three ZSL benchmarks, i.e., CUB, SUN and AWA2, show the superiority or competitiveness of our proposed method against the state-of-the-art ZSL methods. The ablation analysis on real gaze data CUB-VWSW also validates the benefits and accuracy of our gaze estimation module. This work implies the promising benefits of collecting human gaze dataset and automatic gaze estimation algorithms on high-level computer vision tasks. The code is available at https://github.com/osierboy/GEM-ZSL.

cs.CV↗

arXiv2021

Information Bottleneck Constrained Latent Bidirectional Embedding for Zero-Shot Learning

Zero-shot learning (ZSL) aims to recognize novel classes by transferring semantic knowledge from seen classes to unseen classes. Though many ZSL methods rely on a direct mapping between the visual and the semantic space, the calibration deviation and hubness problem limit the generalization capability to unseen classes. Recently emerged generative ZSL methods generate unseen image features to transform ZSL into a supervised classification problem. However, most generative models still suffer from the seen-unseen bias problem as only seen data is used for training. To address these issues, we propose a novel bidirectional embedding based generative model with a tight visual-semantic coupling constraint. We learn a unified latent space that calibrates the embedded parametric distributions of both visual and semantic spaces. Since the embedding from high-dimensional visual features comprise much non-semantic information, the alignment of visual and semantic in latent space would inevitably been deviated. Therefore, we introduce information bottleneck (IB) constraint to ZSL for the first time to preserve essential attribute information during the mapping. Specifically, we utilize the uncertainty estimation and the wake-sleep procedure to alleviate the feature noises and improve model abstraction capability. In addition, our method can be easily extended to transductive ZSL setting by generating labels for unseen images. We then introduce a robust loss to solve this label noise problem. Extensive experimental results show that our method outperforms the state-of-the-art methods in different ZSL settings on most benchmark datasets. The code will be available at https://github.com/osierboy/IBZSL.

cs.CV↗

arXiv2021

Dimensional Control of Octahedral Tilt in SrRuO3 via Infinite-layered Oxides

Manipulation of octahedral distortion at atomic length scale is an effective means to tune the physical ground states of functional oxides. Previous work demonstrates that epitaxial strain and film thickness are variable parameters to modify the octahedral rotation and tilt. However, selective control of bonding geometry by structural propagation from adjacent layers is rarely studied. Here we propose a new route to tune the ferromagnetic response in SrRuO3 (SRO) ultrathin layers by oxygen coordination of adjacent SrCuO2 (SCO) layers. The infinite-layered CuO2 in SCO exhibits a structural transformation from "planar-type" to "chain-type" as reducing film thickness. These two orientations dramatically modify the polyhedral connectivity at the interface, thus altering the octahedral distortion of SRO. The local structural variation changes the spin state of Ru and hybridization strength between Ru 4d and O 2p orbitals, leading to a significant change in the magnetoresistance and anomalous Hall resistivity of SRO layers. These findings could launch further investigations into adaptive control of magnetoelectric properties in quantum oxide heterostructures using oxygen coordination.

cond-mat.mtrl-sci↗

arXiv2021

Near-Room Temperature Ferromagnetic Insulating State in Highly Distorted LaCoO2.5 with CoO5 Square Pyramids

Dedicated control of oxygen vacancies is an important route to functionalizing complex oxide films. It is well-known that tensile strain significantly lowers the oxygen vacancy formation energy, whereas compressive strain plays a minor role. Thus, atomically reconstruction by extracting oxygen from a compressive-strained film is challenging. Here we report an unexpected LaCoO2.5 phase with a zigzag-like oxygen vacancy ordering through annealing a compressive-strained LaCoO3 in vacuum. The synergetic tilt and distortion of CoO5 square pyramids with large La and Co shifts are quantified using scanning transmission electron microscopy. The large in-plane expansion of CoO5 square pyramids weaken the crystal-field splitting and facilitated the ordered high-spin state of Co2+, which produces an insulating ferromagnetic state with a Curie temperature of ~284 K and a saturation magnetization of ~0.25 μB/Co. These results demonstrate that extracting targeted oxygen from a compressive-strained oxide provides an opportunity for creating unexpected crystal structures and novel functionalities.

cond-mat.mtrl-sci↗

arXiv2021

Structural Twinning-induced Insulating Phase in CrN (111) Films

Electronic states of a correlated material can be effectively modified by structural variations delivered from a single-crystal substrate. In this letter, we show that the CrN films grown on MgO (001) substrates have a (001) orientation, whereas the CrN films on α-Al2O3 (0001) substrates are oriented along (111) direction parallel to the surface normal. Transport properties of CrN films are remarkably different depending on crystallographic orientations. The critical thickness for the metal-insulator transition (MIT) in CrN 111 films is significantly larger than that of CrN 001 films. In contrast to CrN 001 films without apparent defects, scanning transmission electron microscopy results reveal that CrN 111 films exhibit strain-induced structural defects, e. g. the periodic horizontal twinning domains, resulting in an increased electron scattering facilitating an insulating state. Understanding the key parameters that determine the electronic properties of ultrathin conductive layers is highly desirable for future technological applications.

cond-mat.mtrl-sci↗

arXiv2021

Realization of AlSb in the double layer honeycomb structure: a robust new class of two-dimensional material

Exploring new two-dimensional (2D) van der Waals (vdW) systems is at the forefront of materials physics. Here, through molecular beam epitaxy on graphene-covered SiC(0001), we report successful growth of AlSb in the double-layer honeycomb (DLHC) structure, a 2D vdW material which has no direct analogue to its 3D bulk and is predicted kinetically stable when freestanding. The structural morphology and electronic structure of the experimental 2D AlSb are characterized with spectroscopic imaging scanning tunneling microscopy and cross-sectional imaging scanning transmission electron microscopy, which compare well to the proposed DLHC structure. The 2D AlSb exhibits a bandgap of 0.93 eV versus the predicted 1.06 eV, which is substantially smaller than the 1.6 eV of bulk. We also attempt the less-stable InSb DLHC structure; however, it grows into bulk islands instead. The successful growth of a DLHC material here opens the door for the realization of a large family of novel 2D DLHC traditional semiconductors with unique excitonic, topological, and electronic properties.

cond-mat.mtrl-sci↗

arXiv2020

Strong Ferromagnetism Achieved via Breathing Lattices in Atomically Thin Cobaltites

Low-dimensional quantum materials that remain strongly ferromagnetic down to mono layer thickness are highly desired for spintronic applications. Although oxide materials are important candidates for next generation of spintronic, ferromagnetism decays severely when the thickness is scaled to the nano meter regime, leading to deterioration of device performance. Here we report a methodology for maintaining strong ferromagnetism in insulating LaCoO3 (LCO) layers down to the thickness of a single unit cell. We find that the magnetic and electronic states of LCO are linked intimately to the structural parameters of adjacent "breathing lattice" SrCuO2 (SCO). As the dimensionality of SCO is reduced, the lattice constant elongates over 10% along the growth direction, leading to a significant distortion of the CoO6 octahedra, and promoting a higher spin state and long-range spin ordering. For atomically thin LCO layers, we observe surprisingly large magnetic moment (0.5 uB/Co) and Curie temperature (75 K), values larger than previously reported for any mono layer oxide. Our results demonstrate a strategy for creating ultra thin ferromagnetic oxides by exploiting atomic hetero interface engineering,confinement-driven structural transformation, and spin-lattice entanglement in strongly correlated materials.

cond-mat.mtrl-sci↗

arXiv2020

Strain-mediated high conductivity in ultrathin antiferromagnetic metallic nitrides

Strain engineering provides the ability to control the ground states and associated phase transition in the epitaxial films. However, the systematic study of intrinsic characters and their strain dependency in transition-metal nitrides remains challenging due to the difficulty in fabricating the stoichiometric and high-quality films. Here we report the observation of electronic state transition in highly crystalline antiferromagnetic CrN films with strain and reduced dimensionality. Shrinking the film thickness to a critical value of ~ 30 unit cells, a profound conductivity reduction accompanied by unexpected volume expansion is observed in CrN films. The electrical conductivity is observed surprisingly when the CrN layer as thin as single unit cell thick, which is far below the critical thickness of most metallic films. We found that the metallicity of an ultrathin CrN film recovers from an insulating behavior upon the removal of as-grown strain by fabrication of first-ever freestanding nitride films. Both first-principles calculations and linear dichroism measurements reveal that the strain-mediated orbital splitting effectively customizes the relatively small bandgap at the Fermi level, leading to exotic phase transition in CrN. The ability to achieve highly conductive nitride ultrathin films by harness strain-controlling over competing phases can be used for utilizing their exceptional characteristics.

cond-mat.mtrl-sci↗

arXiv2020

Beyond Triplet Loss: Person Re-identification with Fine-grained Difference-aware Pairwise Loss

Person Re-IDentification (ReID) aims at re-identifying persons from different viewpoints across multiple cameras. Capturing the fine-grained appearance differences is often the key to accurate person ReID, because many identities can be differentiated only when looking into these fine-grained differences. However, most state-of-the-art person ReID approaches, typically driven by a triplet loss, fail to effectively learn the fine-grained features as they are focused more on differentiating large appearance differences. To address this issue, we introduce a novel pairwise loss function that enables ReID models to learn the fine-grained features by adaptively enforcing an exponential penalization on the images of small differences and a bounded penalization on the images of large differences. The proposed loss is generic and can be used as a plugin to replace the triplet loss to significantly enhance different types of state-of-the-art approaches. Experimental results on four benchmark datasets show that the proposed loss substantially outperforms a number of popular loss functions by large margins; and it also enables significantly improved data efficiency.

cs.CV↗

arXiv2020

The anomalous Hall Effect in a Magnetically Extended Topological Insulator Heterostructure

Constructing heterostructures of a topological insulator (TI) with an undoped magnetic insulator (MI) is a clean and versatile approach to break the time-reversible symmetry in the TI surface states. Despite a lot of efforts, the strength of interfacial magnetic proximity effect (MPE) is still too weak to achieve the quantum anomalous Hall effect and many other topological quantum phenomena. Recently, a new approach based on intercalation of atomic layers of MI, referred to as magnetic extension, was proposed to realize strong MPE [2D Mater. 4, 025082(2017)]. Motivated by this proposal, here, we study a magnetic extension system prepared by molecular beam epitaxy growth of MnSe thin films on topological insulator (Bi,Sb)2Te3. Direct evidence is obtained for intercalation of MnSe atomic layer into a few quintuple layers of (Bi,Sb)2Te3, forming either a double magnetic septuple layer (SL) or an isolated single SL at the interface, where one SL denotes a van der Waals building block consisting of B-A-B-Mn-B-A-B (A=Bi1-xSbx, B= Te1-ySey). The two types of interfaces (namely TI/mono-SL and TI/bi-SL) have different MPE, which is manifested as distinctively different transport behaviors. Specifically, the mono-SL induces a spinflip transition with a sharp change at small magnetic field in the anomalous Hall effect of TI layers, while the bi-SL induces a spin-flop transition with a slow change at large field. Our work demonstrates a useful platform to realize the full potential of the magnetic extension approach for pursuing novel topological physics and related device applications.

physics.app-ph↗