Showing 1–50 of 63 results for author: Baek, H

BankTweak: Adversarial Attack against Multi-Object Trackers by Manipulating Feature Banks

Authors: Woojin Shin, Donghwa Kang, Daejin Choi, Brent Kang, Jinkyu Lee, Hyeongboo Baek

Abstract: Multi-object tracking (MOT) aims to construct moving trajectories for objects, and modern multi-object trackers mainly utilize the tracking-by-detection methodology. Initial approaches to MOT attacks primarily aimed to degrade the detection quality of the frames under attack, thereby reducing accuracy only in those specific frames, highlighting a lack of \textit{efficiency}. To improve efficiency,… ▽ More Multi-object tracking (MOT) aims to construct moving trajectories for objects, and modern multi-object trackers mainly utilize the tracking-by-detection methodology. Initial approaches to MOT attacks primarily aimed to degrade the detection quality of the frames under attack, thereby reducing accuracy only in those specific frames, highlighting a lack of \textit{efficiency}. To improve efficiency, recent advancements manipulate object positions to cause persistent identity (ID) switches during the association phase, even after the attack ends within a few frames. However, these position-manipulating attacks have inherent limitations, as they can be easily counteracted by adjusting distance-related parameters in the association phase, revealing a lack of \textit{robustness}. In this paper, we present \textsf{BankTweak}, a novel adversarial attack designed for MOT trackers, which features efficiency and robustness. \textsf{BankTweak} focuses on the feature extractor in the association phase and reveals vulnerability in the Hungarian matching method used by feature-based MOT systems. Exploiting the vulnerability, \textsf{BankTweak} induces persistent ID switches (addressing \textit{efficiency}) even after the attack ends by strategically injecting altered features into the feature banks without modifying object positions (addressing \textit{robustness}). To demonstrate the applicability, we apply \textsf{BankTweak} to three multi-object trackers (DeepSORT, StrongSORT, and MOTDT) with one-stage, two-stage, anchor-free, and transformer detectors. Extensive experiments on the MOT17 and MOT20 datasets show that our method substantially surpasses existing attacks, exposing the vulnerability of the tracking-by-detection framework to \textsf{BankTweak}. △ Less

Submitted 22 August, 2024; originally announced August 2024.

AT-SNN: Adaptive Tokens for Vision Transformer on Spiking Neural Network

Authors: Donghwa Kang, Youngmoon Lee, Eun-Kyu Lee, Brent Kang, Jinkyu Lee, Hyeongboo Baek

Abstract: In the training and inference of spiking neural networks (SNNs), direct training and lightweight computation methods have been orthogonally developed, aimed at reducing power consumption. However, only a limited number of approaches have applied these two mechanisms simultaneously and failed to fully leverage the advantages of SNN-based vision transformers (ViTs) since they were originally designe… ▽ More In the training and inference of spiking neural networks (SNNs), direct training and lightweight computation methods have been orthogonally developed, aimed at reducing power consumption. However, only a limited number of approaches have applied these two mechanisms simultaneously and failed to fully leverage the advantages of SNN-based vision transformers (ViTs) since they were originally designed for convolutional neural networks (CNNs). In this paper, we propose AT-SNN designed to dynamically adjust the number of tokens processed during inference in SNN-based ViTs with direct training, wherein power consumption is proportional to the number of tokens. We first demonstrate the applicability of adaptive computation time (ACT), previously limited to RNNs and ViTs, to SNN-based ViTs, enhancing it to discard less informative spatial tokens selectively. Also, we propose a new token-merge mechanism that relies on the similarity of tokens, which further reduces the number of tokens while enhancing accuracy. We implement AT-SNN to Spikformer and show the effectiveness of AT-SNN in achieving high energy efficiency and accuracy compared to state-of-the-art approaches on the image classification tasks, CIFAR10, CIFAR-100, and TinyImageNet. For example, our approach uses up to 42.4% fewer tokens than the existing best-performing method on CIFAR-100, while conserving higher accuracy. △ Less

Submitted 22 August, 2024; originally announced August 2024.

Comments: 8 pages

A special subring of the Nagata ring and the Serre's conjecture ring

Authors: Hyungtae Baek, Jung Wook Lim

Abstract: Many ring theorists researched various properties of Nagata rings and Serre's conjecture rings. In this paper, we introduce a subring (refer to the Anderson ring) of both the Nagata ring and the Serre's conjecture ring (up to isomorphism), and investigate properties of the Anderson rings. Additionally, we compare the properties of the Anderson rings with those of Nagata rings and Serre's conjectur… ▽ More Many ring theorists researched various properties of Nagata rings and Serre's conjecture rings. In this paper, we introduce a subring (refer to the Anderson ring) of both the Nagata ring and the Serre's conjecture ring (up to isomorphism), and investigate properties of the Anderson rings. Additionally, we compare the properties of the Anderson rings with those of Nagata rings and Serre's conjecture rings. △ Less

Submitted 19 August, 2024; v1 submitted 16 August, 2024; originally announced August 2024.

Comments: 18 pages

MSC Class: 13A15; 13B25; 13B30

Fast Quantum Convolutional Neural Networks for Low-Complexity Object Detection in Autonomous Driving Applications

Authors: Hankyul Baek, Donghyeon Kim, Joongheon Kim

Abstract: Spurred by consistent advances and innovation in deep learning, object detection applications have become prevalent, particularly in autonomous driving that leverages various visual data. As convolutional neural networks (CNNs) are being optimized, the performances and computation speeds of object detection in autonomous driving have been significantly improved. However, due to the exponentially r… ▽ More Spurred by consistent advances and innovation in deep learning, object detection applications have become prevalent, particularly in autonomous driving that leverages various visual data. As convolutional neural networks (CNNs) are being optimized, the performances and computation speeds of object detection in autonomous driving have been significantly improved. However, due to the exponentially rapid growth in the complexity and scale of data used in object detection, there are limitations in terms of computation speeds while conducting object detection solely with classical computing. Motivated by this, quantum convolution-based object detection (QCOD) is proposed to adopt quantum computing to perform object detection at high speed. The QCOD utilizes our proposed fast quantum convolution that uploads input channel information and re-constructs output channels for achieving reduced computational complexity and thus improving performances. Lastly, the extensive experiments with KITTI autonomous driving object detection dataset verify that the proposed fast quantum convolution and QCOD are successfully operated in real object detection applications. △ Less

Submitted 27 December, 2023; originally announced January 2024.

Comments: 11 pages, 9 figures

On $S$-Noetherian modules and $S$-strong Mori modules

Authors: Hyungtae Baek, Jung Wook Lim

Abstract: In this paper, we study some properties of $S$-Noetherian modules and $S$-strong Mori modules. Among other things, we prove the Hilbert basis theorem for $S$-Noetherian modules and $S$-strong Mori modules. In this paper, we study some properties of $S$-Noetherian modules and $S$-strong Mori modules. Among other things, we prove the Hilbert basis theorem for $S$-Noetherian modules and $S$-strong Mori modules. △ Less

Submitted 19 August, 2024; v1 submitted 2 November, 2023; originally announced November 2023.

Comments: 15 pages

MSC Class: 13A15; 13B25; 13E99

On $S$-$n$-absorbing ideals

Authors: Hyungtae Baek, Hyun Seung Choi, Jung Wook Lim

Abstract: Let $R$ be a commutative ring with identity, $S$ a multiplicative subset of $R$ and $I$ an ideal of $R$ disjoint from $S$. In this paper, we introduce the notion of an $S$-$n$-absorbing ideal which is a generalization of both the $S$-prime ideals and $n$-absorbing ideals. Moreover, we investigate the basic properties, quotient extension, existence and amalgamation of $S$-$n$-absorbing ideals. Let $R$ be a commutative ring with identity, $S$ a multiplicative subset of $R$ and $I$ an ideal of $R$ disjoint from $S$. In this paper, we introduce the notion of an $S$-$n$-absorbing ideal which is a generalization of both the $S$-prime ideals and $n$-absorbing ideals. Moreover, we investigate the basic properties, quotient extension, existence and amalgamation of $S$-$n$-absorbing ideals. △ Less

Submitted 19 August, 2024; v1 submitted 27 October, 2023; originally announced October 2023.

Comments: 18pages

MSC Class: 13A15

Patch-Mix Contrastive Learning with Audio Spectrogram Transformer on Respiratory Sound Classification

Authors: Sangmin Bae, June-Woo Kim, Won-Yang Cho, Hyerim Baek, Soyoun Son, Byungjo Lee, Changwan Ha, Kyongpil Tae, Sungnyun Kim, Se-Young Yun

Abstract: Respiratory sound contains crucial information for the early diagnosis of fatal lung diseases. Since the COVID-19 pandemic, there has been a growing interest in contact-free medical care based on electronic stethoscopes. To this end, cutting-edge deep learning models have been developed to diagnose lung diseases; however, it is still challenging due to the scarcity of medical data. In this study,… ▽ More Respiratory sound contains crucial information for the early diagnosis of fatal lung diseases. Since the COVID-19 pandemic, there has been a growing interest in contact-free medical care based on electronic stethoscopes. To this end, cutting-edge deep learning models have been developed to diagnose lung diseases; however, it is still challenging due to the scarcity of medical data. In this study, we demonstrate that the pretrained model on large-scale visual and audio datasets can be generalized to the respiratory sound classification task. In addition, we introduce a straightforward Patch-Mix augmentation, which randomly mixes patches between different samples, with Audio Spectrogram Transformer (AST). We further propose a novel and effective Patch-Mix Contrastive Learning to distinguish the mixed representations in the latent space. Our method achieves state-of-the-art performance on the ICBHI dataset, outperforming the prior leading score by an improvement of 4.08%. △ Less

Submitted 22 November, 2023; v1 submitted 23 May, 2023; originally announced May 2023.

Comments: INTERSPEECH 2023, Code URL: https://github.com/raymin0223/patch-mix_contrastive_learning

The interplay of field-tunable strongly correlated states in a multi-orbital moiré system

Authors: Aidan J. Campbell, Valerio Vitale, Mauro Brotons-Gisbert, Hyeonjun Baek, Takashi Taniguchi, Kenji Watanabe, Johannes Lischner, Brian D. Gerardot

Abstract: The interplay of charge, spin, lattice, and orbital degrees of freedom leads to a wide range of emergent phenomena in strongly correlated systems. In heterobilayer transition metal dichalcogenide moiré systems, recent observations of Mott insulators and generalized Wigner crystals are well described by triangular lattice single-orbital Hubbard models based on K-valley derived moiré bands. Richer p… ▽ More The interplay of charge, spin, lattice, and orbital degrees of freedom leads to a wide range of emergent phenomena in strongly correlated systems. In heterobilayer transition metal dichalcogenide moiré systems, recent observations of Mott insulators and generalized Wigner crystals are well described by triangular lattice single-orbital Hubbard models based on K-valley derived moiré bands. Richer phase diagrams, mapped onto multi-orbital Hubbard models, are possible with hexagonal lattices in $Γ$-valley derived moiré bands and additional layer degrees of freedom. Here we report the tunable interaction between strongly correlated hole states hosted by $Γ$- and K-derived moiré bands in a monolayer MoSe$_2$ / natural WSe$_2$ bilayer device. To precisely probe the nature of the correlated states, we optically characterise the behaviour of exciton-polarons and distinguish the layer and valley degrees of freedom. We find that the honeycomb $Γ$-band gives rise to a charge-transfer insulator described by a two-orbital Hubbard model with inequivalent $Γ_\mathrm{A}$ and $Γ_\mathrm{B}$ orbitals. With an out-of-plane electric field, we re-order the $Γ_\mathrm{B}$- and K-derived bands energetically, driving an abrupt redistribution of carriers to the layer-polarized K orbital where new correlated states are observed. Finally, by fine-tuning the band-alignment, we obtain degeneracy of the $Γ_\mathrm{B}$ and K orbitals at the Fermi level. In this critical condition, stable Wigner crystals with carriers distributed across the two orbitals are observed until the Fermi-level reaches one hole per lattice site, whereupon the system collapses into a filled $Γ_\mathrm{B}$ orbital. Our results establish a platform to investigate the interplay of charge, spin, lattice, and layer geometry in multi-orbital Hubbard model Hamiltonians. △ Less

Submitted 9 March, 2023; originally announced March 2023.

Comments: 14 pages, 5 figures

Highly Tunable Ground and Excited State Excitonic Dipoles in Multilayer 2H-MoSe$_2$

Authors: Shun Feng, Aidan Campbell, Mauro Brotons-Gisbert, Daniel Andres-Penares, Hyeonjun Baek, Takashi Taniguchi, Kenji Watanabe, Bernhard Urbaszek, Iann C. Gerber, Brian D. Gerardot

Abstract: The fundamental properties of an exciton are determined by the spin, valley, energy, and spatial wavefunctions of the Coulomb bound electron and hole. In van der Waals materials, these attributes can be widely engineered through layer stacking configuration to create highly tunable interlayer excitons with static out-of-plane electric dipoles, at the expense of the strength of the oscillating in-p… ▽ More The fundamental properties of an exciton are determined by the spin, valley, energy, and spatial wavefunctions of the Coulomb bound electron and hole. In van der Waals materials, these attributes can be widely engineered through layer stacking configuration to create highly tunable interlayer excitons with static out-of-plane electric dipoles, at the expense of the strength of the oscillating in-plane dipole responsible for light-matter coupling. Here we show that interlayer excitons in bi- and tri-layer 2H-MoSe$_2$ crystals exhibit electric-field-driven coupling with the ground ($1s$) and excited states ($2s$) of the intralayer A excitons. We demonstrate that the hybrid states of these distinct exciton species provide strong oscillator strength, large permanent dipoles (up to $0.73 \pm 0.01$ enm), high energy tunability (up to $\sim$ 200 meV), and full control of the spin and valley characteristics such that the exciton g-factor can be manipulated over a large range (from -4 to +14). Further, we observe the bi- and tri-layer excited state ($2s$) interlayer excitons and their coupling with the intralayer excitons states ($1s$ and $2s$). Our results, in good agreement with a coupled oscillator model with spin (layer)-selectivity and beyond standard density functional theory calculations, promote multilayer 2H-MoSe$_2$ as a highly tunable platform to explore exciton-exciton interactions with strong light-matter interactions. △ Less

Submitted 19 January, 2023; v1 submitted 29 December, 2022; originally announced December 2022.

Journal ref: Nat. Commun. 15, 4377 (2024)

Quantum Federated Learning with Entanglement Controlled Circuits and Superposition Coding

Authors: Won Joon Yun, Jae Pyoung Kim, Hankyul Baek, Soyi Jung, Jihong Park, Mehdi Bennis, Joongheon Kim

Abstract: While witnessing the noisy intermediate-scale quantum (NISQ) era and beyond, quantum federated learning (QFL) has recently become an emerging field of study. In QFL, each quantum computer or device locally trains its quantum neural network (QNN) with trainable gates, and communicates only these gate parameters over classical channels, without costly quantum communications. Towards enabling QFL und… ▽ More While witnessing the noisy intermediate-scale quantum (NISQ) era and beyond, quantum federated learning (QFL) has recently become an emerging field of study. In QFL, each quantum computer or device locally trains its quantum neural network (QNN) with trainable gates, and communicates only these gate parameters over classical channels, without costly quantum communications. Towards enabling QFL under various channel conditions, in this article we develop a depth-controllable architecture of entangled slimmable quantum neural networks (eSQNNs), and propose an entangled slimmable QFL (eSQFL) that communicates the superposition-coded parameters of eS-QNNs. Compared to the existing depth-fixed QNNs, training the depth-controllable eSQNN architecture is more challenging due to high entanglement entropy and inter-depth interference, which are mitigated by introducing entanglement controlled universal (CU) gates and an inplace fidelity distillation (IPFD) regularizer penalizing inter-depth quantum state differences, respectively. Furthermore, we optimize the superposition coding power allocation by deriving and minimizing the convergence bound of eSQFL. In an image classification task, extensive simulations corroborate the effectiveness of eSQFL in terms of prediction accuracy, fidelity, and entropy compared to Vanilla QFL as well as under different channel conditions and various data distributions. △ Less

Submitted 3 December, 2022; originally announced December 2022.

Quantum Split Neural Network Learning using Cross-Channel Pooling

Authors: Won Joon Yun, Hankyul Baek, Joongheon Kim

Abstract: In recent years, the field of quantum science has attracted significant interest across various disciplines, including quantum machine learning, quantum communication, and quantum computing. Among these emerging areas, quantum federated learning (QFL) has gained particular attention due to the integration of quantum neural networks (QNNs) with traditional federated learning (FL) techniques. In thi… ▽ More In recent years, the field of quantum science has attracted significant interest across various disciplines, including quantum machine learning, quantum communication, and quantum computing. Among these emerging areas, quantum federated learning (QFL) has gained particular attention due to the integration of quantum neural networks (QNNs) with traditional federated learning (FL) techniques. In this study, a novel approach entitled quantum split learning (QSL) is presented, which represents an advanced extension of classical split learning. Previous research in classical computing has demonstrated numerous advantages of split learning, such as accelerated convergence, reduced communication costs, and enhanced privacy protection. To maximize the potential of QSL, cross-channel pooling is introduced, a technique that capitalizes on the distinctive properties of quantum state tomography facilitated by QNNs. Through rigorous numerical analysis, evidence is provided that QSL not only achieves a 1.64\% higher top-1 accuracy compared to QFL but also demonstrates robust privacy preservation in the context of the MNIST classification task. △ Less

Submitted 8 April, 2023; v1 submitted 11 November, 2022; originally announced November 2022.

Neural Architectural Nonlinear Pre-Processing for mmWave Radar-based Human Gesture Perception

Authors: Hankyul Baek, Yoo Jeong, Ha, Minjae Yoo, Soyi Jung, Joongheon Kim

Abstract: In modern on-driving computing environments, many sensors are used for context-aware applications. This paper utilizes two deep learning models, U-Net and EfficientNet, which consist of a convolutional neural network (CNN), to detect hand gestures and remove noise in the Range Doppler Map image that was measured through a millimeter-wave (mmWave) radar. To improve the performance of classification… ▽ More In modern on-driving computing environments, many sensors are used for context-aware applications. This paper utilizes two deep learning models, U-Net and EfficientNet, which consist of a convolutional neural network (CNN), to detect hand gestures and remove noise in the Range Doppler Map image that was measured through a millimeter-wave (mmWave) radar. To improve the performance of classification, accurate pre-processing algorithms are essential. Therefore, a novel pre-processing approach to denoise images before entering the first deep learning model stage increases the accuracy of classification. Thus, this paper proposes a deep neural network based high-performance nonlinear pre-processing method. △ Less

Submitted 7 November, 2022; originally announced November 2022.

Comments: 4 pages, 7 figures

Projection Valued Measure-based Quantum Machine Learning for Multi-Class Classification

Authors: Won Joon Yun, Hankyul Baek, Joongheon Kim

Abstract: In recent years, quantum machine learning (QML) has been actively used for various tasks, e.g., classification, reinforcement learning, and adversarial learning. However, these QML studies are unable to carry out complex tasks due to scalability issues on input and output which is currently the biggest hurdle in QML. Therefore, the purpose of this paper is to overcome the problem of scalability. M… ▽ More In recent years, quantum machine learning (QML) has been actively used for various tasks, e.g., classification, reinforcement learning, and adversarial learning. However, these QML studies are unable to carry out complex tasks due to scalability issues on input and output which is currently the biggest hurdle in QML. Therefore, the purpose of this paper is to overcome the problem of scalability. Motivated by this challenge, we focus on projection-valued measurements (PVM) which utilize the nature of probability amplitude in quantum statistical mechanics. By leveraging PVM, the output dimension is expanded from $q$, which is the number of qubits, to $2^q$. We propose a novel QML framework that utilizes PVM for multi-class classification. Our framework is proven to outperform the state-of-the-art (SOTA) methodologies with various datasets, assuming no more than 6 qubits are used. Furthermore, our PVM-based QML shows about $42.2\%$ better performance than the SOTA framework. △ Less

Submitted 14 November, 2022; v1 submitted 29 October, 2022; originally announced October 2022.

RT-MOT: Confidence-Aware Real-Time Scheduling Framework for Multi-Object Tracking Tasks

Authors: Donghwa Kang, Seunghoon Lee, Hoon Sung Chwa, Seung-Hwan Bae, Chang Mook Kang, Jinkyu Lee, Hyeongboo Baek

Abstract: Different from existing MOT (Multi-Object Tracking) techniques that usually aim at improving tracking accuracy and average FPS, real-time systems such as autonomous vehicles necessitate new requirements of MOT under limited computing resources: (R1) guarantee of timely execution and (R2) high tracking accuracy. In this paper, we propose RT-MOT, a novel system design for multiple MOT tasks, which a… ▽ More Different from existing MOT (Multi-Object Tracking) techniques that usually aim at improving tracking accuracy and average FPS, real-time systems such as autonomous vehicles necessitate new requirements of MOT under limited computing resources: (R1) guarantee of timely execution and (R2) high tracking accuracy. In this paper, we propose RT-MOT, a novel system design for multiple MOT tasks, which addresses R1 and R2. Focusing on multiple choices of a workload pair of detection and association, which are two main components of the tracking-by-detection approach for MOT, we tailor a measure of object confidence for RT-MOT and develop how to estimate the measure for the next frame of each MOT task. By utilizing the estimation, we make it possible to predict tracking accuracy variation according to different workload pairs to be applied to the next frame of an MOT task. Next, we develop a novel confidence-aware real-time scheduling framework, which offers an offline timing guarantee for a set of MOT tasks based on non-preemptive fixed-priority scheduling with the smallest workload pair. At run-time, the framework checks the feasibility of a priority-inversion associated with a larger workload pair, which does not compromise the timing guarantee of every task, and then chooses a feasible scenario that yields the largest tracking accuracy improvement based on the proposed prediction. Our experiment results demonstrate that RT-MOT significantly improves overall tracking accuracy by up to 1.5x, compared to existing popular tracking-by-detection approaches, while guaranteeing timely execution of all MOT tasks. △ Less

Submitted 19 October, 2022; originally announced October 2022.

Comments: Accepted to 2022 Real-Time Systems Symposium (RTSS)

3D Scalable Quantum Convolutional Neural Networks for Point Cloud Data Processing in Classification Applications

Authors: Hankyul Baek, Won Joon Yun, Joongheon Kim

Abstract: With the beginning of the noisy intermediate-scale quantum (NISQ) era, a quantum neural network (QNN) has recently emerged as a solution for several specific problems that classical neural networks cannot solve. Moreover, a quantum convolutional neural network (QCNN) is the quantum-version of CNN because it can process high-dimensional vector inputs in contrast to QNN. However, due to the nature o… ▽ More With the beginning of the noisy intermediate-scale quantum (NISQ) era, a quantum neural network (QNN) has recently emerged as a solution for several specific problems that classical neural networks cannot solve. Moreover, a quantum convolutional neural network (QCNN) is the quantum-version of CNN because it can process high-dimensional vector inputs in contrast to QNN. However, due to the nature of quantum computing, it is difficult to scale up the QCNN to extract a sufficient number of features due to barren plateaus. Motivated by this, a novel 3D scalable QCNN (sQCNN-3D) is proposed for point cloud data processing in classification applications. Furthermore, reverse fidelity training (RF-Train) is additionally considered on top of sQCNN-3D for diversifying features with a limited number of qubits using the fidelity of quantum computing. Our data-intensive performance evaluation verifies that the proposed algorithm achieves desired performance. △ Less

Submitted 18 October, 2022; originally announced October 2022.

Scalable Quantum Convolutional Neural Networks

Authors: Hankyul Baek, Won Joon Yun, Joongheon Kim

Abstract: With the beginning of the noisy intermediate-scale quantum (NISQ) era, quantum neural network (QNN) has recently emerged as a solution for the problems that classical neural networks cannot solve. Moreover, QCNN is attracting attention as the next generation of QNN because it can process high-dimensional vector input. However, due to the nature of quantum computing, it is difficult for the classic… ▽ More With the beginning of the noisy intermediate-scale quantum (NISQ) era, quantum neural network (QNN) has recently emerged as a solution for the problems that classical neural networks cannot solve. Moreover, QCNN is attracting attention as the next generation of QNN because it can process high-dimensional vector input. However, due to the nature of quantum computing, it is difficult for the classical QCNN to extract a sufficient number of features. Motivated by this, we propose a new version of QCNN, named scalable quantum convolutional neural network (sQCNN). In addition, using the fidelity of QC, we propose an sQCNN training algorithm named reverse fidelity training (RF-Train) that maximizes the performance of sQCNN. △ Less

Submitted 11 December, 2022; v1 submitted 25 September, 2022; originally announced September 2022.

FV-Train: Quantum Convolutional Neural Network Training with a Finite Number of Qubits by Extracting Diverse Features

Authors: Hankyul Baek, Won Joon Yun, Joongheon Kim

Abstract: Quantum convolutional neural network (QCNN) has just become as an emerging research topic as we experience the noisy intermediate-scale quantum (NISQ) era and beyond. As convolutional filters in QCNN extract intrinsic feature using quantum-based ansatz, it should use only finite number of qubits to prevent barren plateaus, and it introduces the lack of the feature information. In this paper, we pr… ▽ More Quantum convolutional neural network (QCNN) has just become as an emerging research topic as we experience the noisy intermediate-scale quantum (NISQ) era and beyond. As convolutional filters in QCNN extract intrinsic feature using quantum-based ansatz, it should use only finite number of qubits to prevent barren plateaus, and it introduces the lack of the feature information. In this paper, we propose a novel QCNN training algorithm to optimize feature extraction while using only a finite number of qubits, which is called fidelity-variation training (FV-Training). △ Less

Submitted 18 September, 2022; originally announced September 2022.

Comments: 2 pages, 3 figures

Keypoint based Sign Language Translation without Glosses

Authors: Youngmin Kim, Minji Kwak, Dain Lee, Yeongeun Kim, Hyeongboo Baek

Abstract: Sign Language Translation (SLT) is a task that has not been studied relatively much compared to the study of Sign Language Recognition (SLR). However, the SLR is a study that recognizes the unique grammar of sign language, which is different from the spoken language and has a problem that non-disabled people cannot easily interpret. So, we're going to solve the problem of translating directly spok… ▽ More Sign Language Translation (SLT) is a task that has not been studied relatively much compared to the study of Sign Language Recognition (SLR). However, the SLR is a study that recognizes the unique grammar of sign language, which is different from the spoken language and has a problem that non-disabled people cannot easily interpret. So, we're going to solve the problem of translating directly spoken language in sign language video. To this end, we propose a new keypoint normalization method for performing translation based on the skeleton point of the signer and robustly normalizing these points in sign language translation. It contributed to performance improvement by a customized normalization method depending on the body parts. In addition, we propose a stochastic frame selection method that enables frame augmentation and sampling at the same time. Finally, it is translated into the spoken language through an Attention-based translation model. Our method can be applied to various datasets in a way that can be applied to datasets without glosses. In addition, quantitative experimental evaluation proved the excellence of our method. △ Less

Submitted 13 June, 2022; v1 submitted 22 April, 2022; originally announced April 2022.

Comments: 14 pages, 5 figures

SlimFL: Federated Learning with Superposition Coding over Slimmable Neural Networks

Authors: Won Joon Yun, Yunseok Kwak, Hankyul Baek, Soyi Jung, Mingyue Ji, Mehdi Bennis, Jihong Park, Joongheon Kim

Abstract: Federated learning (FL) is a key enabler for efficient communication and computing, leveraging devices' distributed computing capabilities. However, applying FL in practice is challenging due to the local devices' heterogeneous energy, wireless channel conditions, and non-independently and identically distributed (non-IID) data distributions. To cope with these issues, this paper proposes a novel… ▽ More Federated learning (FL) is a key enabler for efficient communication and computing, leveraging devices' distributed computing capabilities. However, applying FL in practice is challenging due to the local devices' heterogeneous energy, wireless channel conditions, and non-independently and identically distributed (non-IID) data distributions. To cope with these issues, this paper proposes a novel learning framework by integrating FL and width-adjustable slimmable neural networks (SNN). Integrating FL with SNNs is challenging due to time-varying channel conditions and data distributions. In addition, existing multi-width SNN training algorithms are sensitive to the data distributions across devices, which makes SNN ill-suited for FL. Motivated by this, we propose a communication and energy-efficient SNN-based FL (named SlimFL) that jointly utilizes superposition coding (SC) for global model aggregation and superposition training (ST) for updating local models. By applying SC, SlimFL exchanges the superposition of multiple-width configurations decoded as many times as possible for a given communication throughput. Leveraging ST, SlimFL aligns the forward propagation of different width configurations while avoiding inter-width interference during backpropagation. We formally prove the convergence of SlimFL. The result reveals that SlimFL is not only communication-efficient but also deals with non-IID data distributions and poor channel conditions, which is also corroborated by data-intensive simulations. △ Less

Submitted 21 December, 2022; v1 submitted 26 March, 2022; originally announced March 2022.

Comments: arXiv admin note: text overlap with arXiv:2112.02543

Exciton-polarons in the presence of strongly correlated electronic states in a MoSe$_2$/WSe$_2$ moiré superlattice

Authors: Aidan J. Campbell, Mauro Brotons-Gisbert, Hyeonjun Baek, Valerio Vitale, Takashi Taniguchi, Kenji Watanabe, Johannes Lischner, Brian D. Gerardot

Abstract: Two-dimensional moiré materials provide a highly tunable platform to investigate strongly correlated electronic states. Such emergent many-body phenomena can be optically probed in moiré systems created by stacking two layers of transition metal dichalcogenide semiconductors: optically injected excitons can interact with itinerant carriers occupying narrow moiré bands to form exciton-polarons sens… ▽ More Two-dimensional moiré materials provide a highly tunable platform to investigate strongly correlated electronic states. Such emergent many-body phenomena can be optically probed in moiré systems created by stacking two layers of transition metal dichalcogenide semiconductors: optically injected excitons can interact with itinerant carriers occupying narrow moiré bands to form exciton-polarons sensitive to strong correlations. Here, we investigate the behaviour of excitons dressed by a Fermi sea localised by the moiré superlattice of a molybdenum diselenide (MoSe$_2$) / tungsten diselenide (WSe$_2$) twisted hetero-bilayer. At a multitude of fractional fillings of the moiré lattice, we observe ordering of both electrons and holes into stable correlated electronic states. Magneto-optical measurements reveal extraordinary Zeeman splittings of the exciton-polarons due to exchange interactions in the correlated hole phases, with a maximum close to the correlated state at one hole per site. The temperature dependence of the Zeeman splitting reveals antiferromagnetic ordering of the correlated holes across a wide range of fractional fillings. Our results illustrate the nature of exciton-polarons in the presence of strongly correlated electronic states and reveal the rich potential of the MoSe$_2$/WSe$_2$ platform for investigations of Fermi-Hubbard and Bose-Hubbard physics. △ Less

Submitted 28 April, 2022; v1 submitted 17 February, 2022; originally announced February 2022.

Comments: 10 pages, 4 figures

Visualization of reaction chemistry in W-KClO4-BaCrO4 delay mixtures via a Sestak-Berggren model based isoconversional method

Authors: Youngjoon Han, Soyeon Kim, Soyun Joo, Chungik Oh, Hojun Lee, Chi Hao Liow, Moon Soo Park, Dong Hyeon Baek, Seungbum Hong

Abstract: The combustion delay mixture of tungsten (W), potassium perchlorate (KClO4), and barium chromate (BaCrO4), also known as the WKB mixture, has long been considered to be an integral part of military-grade ammunition. Despite its long history, however, their progressive reaction dynamics remains a question mark, especially due to the complex nature of their combustion reaction. As opposed to a one-s… ▽ More The combustion delay mixture of tungsten (W), potassium perchlorate (KClO4), and barium chromate (BaCrO4), also known as the WKB mixture, has long been considered to be an integral part of military-grade ammunition. Despite its long history, however, their progressive reaction dynamics remains a question mark, especially due to the complex nature of their combustion reaction. As opposed to a one-step oxidation commonly observed in conventional combustions, the WKB mixture is associated with a multibody reaction between its solid-state components. To this end, the emergence of three combustion peaks, which we corresponded with disparate chemical reactions, was observed using thermogravimetric analysis on two separate WKB mixtures with differing mixture ratios. We applied the stepwise isoconversional method on each of the peaks to match the combustion chemistry it represents to the Sestak-Berggren model and computed the conceptual activation energy. Further plotting the logarithmic pre-exponential factor as a function of the reaction progress, we demonstrate a method of using the plot as an intuitive tool to understand the dynamics of individual reactions that compose multi-step chemical reactions. Our study provides a systematic approach in visualizing the reaction chemistry, thereby strengthening the analytical arsenal against reaction dynamics of combustion compounds in general. △ Less

Submitted 10 December, 2021; originally announced December 2021.

Comments: 10 pages, 5 figures, and 1 table

Communication and Energy Efficient Slimmable Federated Learning via Superposition Coding and Successive Decoding

Authors: Hankyul Baek, Won Joon Yun, Soyi Jung, Jihong Park, Mingyue Ji, Joongheon Kim, Mehdi Bennis

Abstract: Mobile devices are indispensable sources of big data. Federated learning (FL) has a great potential in exploiting these private data by exchanging locally trained models instead of their raw data. However, mobile devices are often energy limited and wirelessly connected, and FL cannot cope flexibly with their heterogeneous and time-varying energy capacity and communication throughput, limiting the… ▽ More Mobile devices are indispensable sources of big data. Federated learning (FL) has a great potential in exploiting these private data by exchanging locally trained models instead of their raw data. However, mobile devices are often energy limited and wirelessly connected, and FL cannot cope flexibly with their heterogeneous and time-varying energy capacity and communication throughput, limiting the adoption. Motivated by these issues, we propose a novel energy and communication efficient FL framework, coined SlimFL. To resolve the heterogeneous energy capacity problem, each device in SlimFL runs a width-adjustable slimmable neural network (SNN). To address the heterogeneous communication throughput problem, each full-width (1.0x) SNN model and its half-width ($0.5$x) model are superposition-coded before transmission, and successively decoded after reception as the 0.5x or $1.0$x model depending on the channel quality. Simulation results show that SlimFL can simultaneously train both $0.5$x and $1.0$x models with reasonable accuracy and convergence speed, compared to its vanilla FL counterpart separately training the two models using $2$x more communication resources. Surprisingly, SlimFL achieves even higher accuracy with lower energy footprints than vanilla FL for poor channels and non-IID data distributions, under which vanilla FL converges slowly. △ Less

Submitted 5 December, 2021; originally announced December 2021.

Comments: 11 pages, 10 Figures, presented at the International Workshop on Federated Learning for User Privacy and Data Confidentiality in Conjunction with ICML 2021 (FL-ICML'21). arXiv admin note: substantial text overlap with arXiv:2112.02543

Joint Superposition Coding and Training for Federated Learning over Multi-Width Neural Networks

Authors: Hankyul Baek, Won Joon Yun, Yunseok Kwak, Soyi Jung, Mingyue Ji, Mehdi Bennis, Jihong Park, Joongheon Kim

Abstract: This paper aims to integrate two synergetic technologies, federated learning (FL) and width-adjustable slimmable neural network (SNN) architectures. FL preserves data privacy by exchanging the locally trained models of mobile devices. By adopting SNNs as local models, FL can flexibly cope with the time-varying energy capacities of mobile devices. Combining FL and SNNs is however non-trivial, parti… ▽ More This paper aims to integrate two synergetic technologies, federated learning (FL) and width-adjustable slimmable neural network (SNN) architectures. FL preserves data privacy by exchanging the locally trained models of mobile devices. By adopting SNNs as local models, FL can flexibly cope with the time-varying energy capacities of mobile devices. Combining FL and SNNs is however non-trivial, particularly under wireless connections with time-varying channel conditions. Furthermore, existing multi-width SNN training algorithms are sensitive to the data distributions across devices, so are ill-suited to FL. Motivated by this, we propose a communication and energy-efficient SNN-based FL (named SlimFL) that jointly utilizes superposition coding (SC) for global model aggregation and superposition training (ST) for updating local models. By applying SC, SlimFL exchanges the superposition of multiple width configurations that are decoded as many as possible for a given communication throughput. Leveraging ST, SlimFL aligns the forward propagation of different width configurations, while avoiding the inter-width interference during backpropagation. We formally prove the convergence of SlimFL. The result reveals that SlimFL is not only communication-efficient but also can counteract non-IID data distributions and poor channel conditions, which is also corroborated by simulations. △ Less

Submitted 5 December, 2021; originally announced December 2021.

Comments: 10 pages, 7 figures, Accepted to IEEE INFOCOM 2022

Electrically driven strain-induced deterministic single-photon emitters in a van der Waals heterostructure

Authors: Jae-Pil So, Ha-Reem Kim, Hyeonjun Baek, Hoo-Cheol Lee, Woong Huh, Yoon Seok Kim, Kenji Watanabe, Takashi Taniguchi, Jungkil Kim, Chul-Ho Lee, Hong-Gyu Park

Abstract: Quantum confinement in atomically-thin TMDCs enables the realization of deterministic single-photon emitters. The position and polarization control of single photons have been achieved via local strain engineering using nanostructures. However, most existing TMDC-based emitters are operated by optical pumping, while the emission sites in electrically pumped emitters are uncontrolled. Here, we demo… ▽ More Quantum confinement in atomically-thin TMDCs enables the realization of deterministic single-photon emitters. The position and polarization control of single photons have been achieved via local strain engineering using nanostructures. However, most existing TMDC-based emitters are operated by optical pumping, while the emission sites in electrically pumped emitters are uncontrolled. Here, we demonstrate electrically driven single-photon emitters located at the positions where strains are induced by atomic-force-microscope indentation on a van der Waals heterostructure consisting of graphene, hexagonal-boron nitride, and tungsten diselenide. The optical, electrical, and mechanical properties induced by the local strain gradient were systematically analyzed. In particular, single-photon emission was observed at the indentation sites at 4 K. The emission exhibits photon anti-bunching behavior with a g(2)(0) value of ~0.3, intensity saturation and a linearly cross-polarized doublet. This robust spatial control of electrically driven single-photon emitters will pave the way for the practical implementation of integrated quantum light sources. △ Less

Submitted 4 May, 2021; originally announced May 2021.

Comments: 29 pages, 8 figures

Optical read-out of Coulomb staircases in a moiré superlattice via trapped interlayer trions

Authors: Hyeonjun Baek, Mauro Brotons-Gisbert, Aidan Campbell, Valerio Vitale, Johannes Lischner, Kenji Watanabe, Takashi Taniguchi, Brian D. Gerardot

Abstract: Moiré patterns with a superlattice potential can be formed by vertically stacking two layered materials with a relative twist or lattice constant mismatch. The moiré superlattice can generate flat bands that result in new correlated insulating, superconducting, and topological states. Strong electron correlations, tunable by the fractional filling, have been observed in both graphene and transitio… ▽ More Moiré patterns with a superlattice potential can be formed by vertically stacking two layered materials with a relative twist or lattice constant mismatch. The moiré superlattice can generate flat bands that result in new correlated insulating, superconducting, and topological states. Strong electron correlations, tunable by the fractional filling, have been observed in both graphene and transition metal dichalcogenide (TMD) based systems. In addition, in TMD based systems, the moiré potential landscape can trap interlayer excitons (IX) at specific atomic registries. Here we report that spatially isolated trapped IX in a molybdenum diselenide/tungsten diselenide heterobilayer device provide a sensitive optical probe of carrier filling in their immediate environment. By mapping the spatial positions of individual trapped IX, we are able to spectrally track the emitters as the moiré lattice is filled with excess carriers. Upon initial doping of the heterobilayer, neutral trapped IX form charged IX (IX trions) uniformly with a binding energy of ~7 meV. Upon further doping, the empty superlattice sites sequentially fill, creating a Coulomb staircase: stepwise changes in the IX trion emission energy due to Coulomb interactions with carriers at nearest neighbour moiré sites. This non-invasive, highly local technique can complement transport and non-local optical sensing techniques to characterise Coulomb interaction energies, visualise charge correlated states, or probe local disorder in a moiré superlattice. △ Less

Submitted 29 July, 2021; v1 submitted 2 February, 2021; originally announced February 2021.

Comments: 10 pages, 5 figures

Moiré-trapped interlayer trions in a charge-tunable WSe$_2$/MoSe$_2$ heterobilayer

Authors: Mauro Brotons-Gisbert, Hyeonjun Baek, Aidan Campbell, Kenji Watanabe, Takashi Taniguchi, Brian D. Gerardot

Abstract: Transition metal dichalcogenide heterobilayers offer attractive opportunities to realize lattices of interacting bosons with several degrees of freedom. Such heterobilayers can feature moiré patterns that modulate their electronic band structure, leading to spatial confinement of single interlayer excitons (IXs) that act as quantum emitters with $C_3$ symmetry. However, the narrow emission linewid… ▽ More Transition metal dichalcogenide heterobilayers offer attractive opportunities to realize lattices of interacting bosons with several degrees of freedom. Such heterobilayers can feature moiré patterns that modulate their electronic band structure, leading to spatial confinement of single interlayer excitons (IXs) that act as quantum emitters with $C_3$ symmetry. However, the narrow emission linewidths of the quantum emitters contrast with a broad ensemble IX emission observed in nominally identical heterobilayers, opening a debate regarding the origin of IX emission. Here we report the continuous evolution from a few trapped IXs to an ensemble of IXs with both triplet and singlet spin configurations in a gate-tunable $2H$-MoSe$_2$/WSe$_2$ heterobilayer. We observe signatures of dipolar interactions in the IX ensemble regime which, when combined with magneto-optical spectroscopy, reveal that the narrow quantum-dot-like and broad ensemble emission originate from IXs trapped in moiré potentials with the same atomic registry. Finally, electron doping leads to the formation of three different species of localised negative trions with contrasting spin-valley configurations, among which we observe both intervalley and intravalley IX trions with spin-triplet optical transitions. Our results identify the origin of IX emission in MoSe$_2$/WSe$_2$ heterobilayers and highlight the important role of exciton-exciton interactions and Fermi-level control in these highly tunable quantum materials. △ Less

Submitted 13 August, 2021; v1 submitted 19 January, 2021; originally announced January 2021.

Comments: 12 pages, 4 figures

Journal ref: Phys. Rev. X 11, 031033 (2021)

Automated 3D cephalometric landmark identification using computerized tomography

Authors: Hye Sun Yun, Chang Min Hyun, Seong Hyeon Baek, Sang-Hwy Lee, Jin Keun Seo

Abstract: Identification of 3D cephalometric landmarks that serve as proxy to the shape of human skull is the fundamental step in cephalometric analysis. Since manual landmarking from 3D computed tomography (CT) images is a cumbersome task even for the trained experts, automatic 3D landmark detection system is in a great need. Recently, automatic landmarking of 2D cephalograms using deep learning (DL) has a… ▽ More Identification of 3D cephalometric landmarks that serve as proxy to the shape of human skull is the fundamental step in cephalometric analysis. Since manual landmarking from 3D computed tomography (CT) images is a cumbersome task even for the trained experts, automatic 3D landmark detection system is in a great need. Recently, automatic landmarking of 2D cephalograms using deep learning (DL) has achieved great success, but 3D landmarking for more than 80 landmarks has not yet reached a satisfactory level, because of the factors hindering machine learning such as the high dimensionality of the input data and limited amount of training data due to ethical restrictions on the use of medical data. This paper presents a semi-supervised DL method for 3D landmarking that takes advantage of anonymized landmark dataset with paired CT data being removed. The proposed method first detects a small number of easy-to-find reference landmarks, then uses them to provide a rough estimation of the entire landmarks by utilizing the low dimensional representation learned by variational autoencoder (VAE). Anonymized landmark dataset is used for training the VAE. Finally, coarse-to-fine detection is applied to the small bounding box provided by rough estimation, using separate strategies suitable for mandible and cranium. For mandibular landmarks, patch-based 3D CNN is applied to the segmented image of the mandible (separated from the maxilla), in order to capture 3D morphological features of mandible associated with the landmarks. We detect 6 landmarks around the condyle all at once, instead of one by one, because they are closely related to each other. For cranial landmarks, we again use VAE-based latent representation for more accurate annotation. In our experiment, the proposed method achieved an averaged 3D point-to-point error of 2.91 mm for 90 landmarks only with 15 paired training data. △ Less

Submitted 16 December, 2020; originally announced January 2021.

Breaking Moravec's Paradox: Visual-Based Distribution in Smart Fashion Retail

Authors: Shin Woong Sung, Hyunsuk Baek, Hyeonjun Sim, Eun Hie Kim, Hyunwoo Hwangbo, Young Jae Jang

Abstract: In this paper, we report an industry-academia collaborative study on the distribution method of fashion products using an artificial intelligence (AI) technique combined with an optimization method. To meet the current fashion trend of short product lifetimes and an increasing variety of styles, the company produces limited volumes of a large variety of styles. However, due to the limited volume o… ▽ More In this paper, we report an industry-academia collaborative study on the distribution method of fashion products using an artificial intelligence (AI) technique combined with an optimization method. To meet the current fashion trend of short product lifetimes and an increasing variety of styles, the company produces limited volumes of a large variety of styles. However, due to the limited volume of each style, some styles may not be distributed to some off-line stores. As a result, this high-variety, low-volume strategy presents another challenge to distribution managers. We collaborated with KOLON F/C, one of the largest fashion business units in South Korea, to develop models and an algorithm to optimally distribute the products to the stores based on the visual images of the products. The team developed a deep learning model that effectively represents the styles of clothes based on their visual image. Moreover, the team created an optimization model that effectively determines the product mix for each store based on the image representation of clothes. In the past, computers were only considered to be useful for conducting logical calculations, and visual perception and cognition were considered to be difficult computational tasks. The proposed approach is significant in that it uses both AI (perception and cognition) and mathematical optimization (logical calculation) to address a practical supply chain problem, which is why the study was called "Breaking Moravec's Paradox." △ Less

Submitted 9 July, 2020; originally announced July 2020.

Comments: 10 pages, 19 figures, The fifth international workshop on fashion and KDD, KDD 2020

Highly tunable quantum light from moiré trapped excitons

Authors: Hyeonjun Baek, Mauro Brotons-Gisbert, Zhe Xian Koong, Aidan Campbell, Markus Rambach, Kenji Watanabe, Takashi Taniguchi, Brian D. Gerardot

Abstract: Photon antibunching, a hallmark of quantum light, has been observed in the correlations of light from isolated atomic and atomic-like solid-state systems. Two-dimensional semiconductor heterostructures offer a unique method to create a quantum light source: a small lattice mismatch or relative twist in a heterobilayer can create moiré trapping potentials for excitons which are predicted to create… ▽ More Photon antibunching, a hallmark of quantum light, has been observed in the correlations of light from isolated atomic and atomic-like solid-state systems. Two-dimensional semiconductor heterostructures offer a unique method to create a quantum light source: a small lattice mismatch or relative twist in a heterobilayer can create moiré trapping potentials for excitons which are predicted to create arrays of quantum emitters. While signatures of moiré trapped excitons have been observed, their quantum nature has yet to be confirmed. Here we report photon antibunching from single moiré trapped interlayer excitons in a heterobilayer. Via polarization resolved magneto-optical spectroscopy, we demonstrate the discrete anharmonic spectra arise from bound band-edge electron-hole pairs trapped in moiré potentials. Finally, using an out-of-plane electric field, we exploit the large permanent dipole of interlayer excitons to achieve large DC Stark tuning, up to 40 meV, of the quantum emitters. Our results confirm the quantum nature of moiré confined excitons and open opportunities to investigate their inhomogeneity and interactions between the emitters or tune single emitters into resonance with cavity modes or other emitters. △ Less

Submitted 15 January, 2020; v1 submitted 13 January, 2020; originally announced January 2020.

Comments: 8 pages, 4 figures (for main manuscript)

Deep Learning-Based Solvability of Underdetermined Inverse Problems in Medical Imaging

Authors: Chang Min Hyun, Seong Hyeon Baek, Mingyu Lee, Sung Min Lee, Jin Keun Seo

Abstract: Recently, with the significant developments in deep learning techniques, solving underdetermined inverse problems has become one of the major concerns in the medical imaging domain. Typical examples include undersampled magnetic resonance imaging, interior tomography, and sparse-view computed tomography, where deep learning techniques have achieved excellent performances. Although deep learning me… ▽ More Recently, with the significant developments in deep learning techniques, solving underdetermined inverse problems has become one of the major concerns in the medical imaging domain. Typical examples include undersampled magnetic resonance imaging, interior tomography, and sparse-view computed tomography, where deep learning techniques have achieved excellent performances. Although deep learning methods appear to overcome the limitations of existing mathematical methods when handling various underdetermined problems, there is a lack of rigorous mathematical foundations that would allow us to elucidate the reasons for the remarkable performance of deep learning methods. This study focuses on learning the causal relationship regarding the structure of the training data suitable for deep learning, to solve highly underdetermined inverse problems. We observe that a majority of the problems of solving underdetermined linear systems in medical imaging are highly non-linear. Furthermore, we analyze if a desired reconstruction map can be learnable from the training data and underdetermined system. △ Less

Submitted 25 June, 2020; v1 submitted 6 January, 2020; originally announced January 2020.

Spin-layer locking of interlayer excitons trapped in moiré potentials

Authors: Mauro Brotons-Gisbert, Hyeonjun Baek, Alejandro Molina-Sánchez, Aidan Campbell, Eleanor Scerri, Daniel White, Kenji Watanabe, Takashi Taniguchi, Cristian Bonato, Brian D. Gerardot

Abstract: Van der Waals heterostructures offer attractive opportunities to design quantum materials. For instance, transition metal dichalcogenides (TMDs) possess three quantum degrees of freedom: spin, valley index, and layer index. Further, twisted TMD heterobilayers can form moiré patterns that modulate the electronic band structure according to atomic registry, leading to spatial confinement of interlay… ▽ More Van der Waals heterostructures offer attractive opportunities to design quantum materials. For instance, transition metal dichalcogenides (TMDs) possess three quantum degrees of freedom: spin, valley index, and layer index. Further, twisted TMD heterobilayers can form moiré patterns that modulate the electronic band structure according to atomic registry, leading to spatial confinement of interlayer exciton (IXs). Here we report the observation of spin-layer locking of IXs trapped in moiré potentials formed in a heterostructure of bilayer 2H-MoSe$_2$ and monolayer WSe$_2$. The phenomenon of locked electron spin and layer index leads to two quantum-confined IX species with distinct spin-layer-valley configurations. Furthermore, we observe that the atomic registries of the moiré trapping sites in the three layers are intrinsically locked together due to the 2H-type stacking characteristic of bilayer TMDs. These results identify the layer index as a useful degree of freedom to engineer tunable few-level quantum systems in two-dimensional heterostructures. △ Less

Submitted 2 June, 2020; v1 submitted 10 August, 2019; originally announced August 2019.

Comments: 10 pages, 4 figures. Supplementary Material available at: https://www.dropbox.com/s/8331cfoasdofwkq/Brotons-Gisbert_Suppl_Info.pdf?dl=0

Journal ref: Nature Materials 19, 630-636 (2020)

Direct Band Gap Semiconducting Holey Graphyne: Structure, Synthesis and Potential Applications

Authors: Xinghui Liu1, Soo Min Cho, Shiru Lin, Eunbhin Yun, Eun Hee Baek, Zhongfang Chen, Do Hyun Ryu, Hyoyoung Lee

Abstract: Here we report two-dimensional (2D) single-crystalline holey-graphyne (HGY) created an interfacial two-solvent system through a Castro-Stephens coupling reaction from 1,3,5-tribromo-2,4,6-triethynylbenzene. HGY is a new type of 2D carbon allotrope whose structure is comprised of a pattern of six-vertex and eight-vertex rings. The carbon-carbon 2D network of HGY is alternately linked between benzen… ▽ More Here we report two-dimensional (2D) single-crystalline holey-graphyne (HGY) created an interfacial two-solvent system through a Castro-Stephens coupling reaction from 1,3,5-tribromo-2,4,6-triethynylbenzene. HGY is a new type of 2D carbon allotrope whose structure is comprised of a pattern of six-vertex and eight-vertex rings. The carbon-carbon 2D network of HGY is alternately linked between benzene rings and sp (carbon-carbon triple bond) bonding. The ratio of the sp over sp2 bonding is 50%. It is confirmed that HGY is stable by DFT calculation. The vibrational, optic, and electric properties of HGY are investigated theoretically and experimentally. It is a p-type semiconductor that embraces a natural direct band gap (~ 1.0 eV) with high hole mobility and electron mobility at room temperature. This report is expected to help develop a new types of carbon-based semiconductor devices with high mobility. △ Less

Submitted 8 July, 2019; originally announced July 2019.

Tuning the interplay between nematicity and spin fluctuations in Na$_{1-x}$Li$_x$FeAs superconductors

Authors: S. H. Baek, Dilip Bhoi, Woohyun Nam, Bumsung Lee, D. V. Efremov, B. Büchner, Kee Hoon Kim

Abstract: Strong interplay of spin and charge/orbital degrees of freedom is the fundamental characteristic of the iron-based superconductors (FeSCs), which leads to the emergence of a nematic state as a rule in the vicinity of the antiferromagnetic state. Despite intense debate for many years, however, whether nematicity is driven by spin or orbital fluctuations remains unsettled. Here, by use of transport,… ▽ More Strong interplay of spin and charge/orbital degrees of freedom is the fundamental characteristic of the iron-based superconductors (FeSCs), which leads to the emergence of a nematic state as a rule in the vicinity of the antiferromagnetic state. Despite intense debate for many years, however, whether nematicity is driven by spin or orbital fluctuations remains unsettled. Here, by use of transport, magnetization, and $^{75}$As nuclear magnetic resonance (NMR) measurements, we show a striking transformation of the relationship between nematicity and spin fluctuations (SFs) in Na$_{1-x}$Li$_x$FeAs; For $x\leq 0.02$, the nematic transition promotes SFs. In contrast, for $x\geq 0.03$, the system undergoes a non-magnetic phase transition at a temperature $T_0$ into a distinct nematic state that suppresses SFs. Such a drastic change of the spin fluctuation spectrum associated with nematicity by small doping is highly unusual, and provides insights into the origin and nature of nematicity in FeSCs. △ Less

Submitted 30 May, 2018; originally announced May 2018.

Comments: 24 pages, 5 figures, 6 supplementary figures; Published in Nature Communications

Journal ref: Nature Communications 9, 2139 (2018)

Computations of atom spectra

Authors: Rune Harder Bak, Henrik Holm

Abstract: This is a contribution to the theory of atoms in abelian categories recently developed in a series of papers by Kanda. We present a method that enables one to explicitly compute the atom spectrum of the module category over a wide range of non-commutative rings. We illustrate our method and results by several examples. This is a contribution to the theory of atoms in abelian categories recently developed in a series of papers by Kanda. We present a method that enables one to explicitly compute the atom spectrum of the module category over a wide range of non-commutative rings. We illustrate our method and results by several examples. △ Less

Submitted 25 September, 2018; v1 submitted 11 May, 2018; originally announced May 2018.

Comments: 16 pages. Final version to appear in Mathematische Nachrichten

MSC Class: 16G20; 18E10

Direct Limit closure of induced Quiver Representations

Authors: Rune Harder Bak

Abstract: In 2004 and 2005 Enochs et al. characterized the flat and projective quiver-representations of left rooted quivers. The proofs can be understood as filtering the classes $Φ(\operatorname{Add}\mathscr X)$ and $Φ(\varinjlim\mathscr X)$ when $\mathscr X$ is the finitely generated projective modules over a ring. In this paper we generalize the above and show that $Φ(\mathscr X)$ can always be filtered… ▽ More In 2004 and 2005 Enochs et al. characterized the flat and projective quiver-representations of left rooted quivers. The proofs can be understood as filtering the classes $Φ(\operatorname{Add}\mathscr X)$ and $Φ(\varinjlim\mathscr X)$ when $\mathscr X$ is the finitely generated projective modules over a ring. In this paper we generalize the above and show that $Φ(\mathscr X)$ can always be filtered for any class $\mathscr X$ in any AB5-abelian category. With an emphasis on $Φ(\varinjlim\mathscr X)$ we investigate the Gorenstein homological situation. Using an abstract version of Pontryagin duals in abelian categories we give a more general characterization of the flat representations and end up by describing the Gorenstein flat quiver representations over right coherent rings. △ Less

Submitted 10 May, 2018; originally announced May 2018.

Comments: 21 pages

MSC Class: 18E10; 16G20; 18A30

Integer and Fractional Quantum Hall effect in Ultra-high Quality Few-layer Black Phosphorus Transistors

Authors: Jiawei Yang, Son Tran, Jason Wu, Shi Che, Petr Stepanov, Takashi Taniguchi, Kenji Watanabe, Hongwoo Baek, Dmitry Smirnov, Ruoyu Chen, Chun Ning Lau

Abstract: As a high mobility two-dimensional semiconductor with strong structural and electronic anisotropy, atomically thin black phosphorus (BP) provides a new playground for investigating the quantum Hall (QH) effect, including outstanding questions such as the functional dependence of Landau level (LL) gaps on magnetic field B, and possible anisotropic fractional QH states. Using encapsulating few-layer… ▽ More As a high mobility two-dimensional semiconductor with strong structural and electronic anisotropy, atomically thin black phosphorus (BP) provides a new playground for investigating the quantum Hall (QH) effect, including outstanding questions such as the functional dependence of Landau level (LL) gaps on magnetic field B, and possible anisotropic fractional QH states. Using encapsulating few-layer BP transistors with mobility up to 55,000 cm2/Vs, we extract LL gaps over an exceptionally wide range of B for QH states at filling factors ν=-1 to -4, which are determined to be linear in B, thus resolving a controversy raised by its anisotropy. Furthermore, a fractional QH state at ν~ -4/3 and an additional feature at -0.56+/- 0.1 are observed, underscoring BP as a tunable 2D platform for exploring electron interactions. △ Less

Submitted 6 April, 2018; originally announced April 2018.

Journal ref: Nano Lett. 18, 1, 229-234 (2018)

THz-Pump and X-Ray-Probe Sources Based on an Electron Linac

Authors: Sadiq Setiniyaz, Seong Hee Park, Hyun Woo Kim, Nikolay A. Vinokurov, Kyu-Ha Jang, Kitae Lee, In Hyung Baek, Young Uk Jeong

Abstract: We describe a compact THz-pump and X-ray-probe beamline, based on an electron linac, for ultrafast time-resolved diffraction applications. Two high-energy electron ($γ>50$) bunches, $5~$ns apart, impinge upon a single-foil or a multifoil radiator and generate THz radiation and X-rays simultaneously. The THz pulse from the first bunch is synchronized to the X-ray beam of the second bunch by using a… ▽ More We describe a compact THz-pump and X-ray-probe beamline, based on an electron linac, for ultrafast time-resolved diffraction applications. Two high-energy electron ($γ>50$) bunches, $5~$ns apart, impinge upon a single-foil or a multifoil radiator and generate THz radiation and X-rays simultaneously. The THz pulse from the first bunch is synchronized to the X-ray beam of the second bunch by using an adjustable optical delay of THz pulse. The peak power of THz radiation from the multifoil radiator is estimated to be $0.14~$GW for a $200~$pC well-optimized electron bunch. GEANT4 simulations show a carbon foil with thickness of $0.5~-~1.0~$mm has the highest yield of $10~-~20~$keV hard X-rays for a $25~$MeV beam, which is approximately $10^3$ photons/(keV pC-electrons) within a few degrees of the polar angle. A carbon multifoil radiator with $35$ foils ($25~$$μm~$thick each) can generate close to $10^3$ hard X-rays/(keV pC-electrons) within a $2^\circ$ acceptance angle. With $200~$pC charge and $100~$Hz repetition rate, we can generate $10^7$ X-rays per $1~$keV energy bin per second or $10^5$ X-rays per $1~$keV energy bin per pulse. The longitudinal time profile of X-ray pulse ranges from $400~-~600~$fs depending on the acceptance angle. The broadening of the time duration of X-ray pulse is observed owing to its diverging effect. A double-crystal monochromator (DCM) will be used to select and transport the desired X-rays to the sample. The heating of the radiators by an electron beam is negligible because of the low beam current. △ Less

Submitted 20 November, 2017; originally announced November 2017.

Comments: 10 pages, 17 figures

Journal ref: Review of Scientific Instruments 88, 113306 (2017)

Strain Engineering a $4a\times\sqrt{3}a$ Charge Density Wave Phase in Transition Metal Dichalcogenide 1T-VSe$_2$

Authors: Duming Zhang, Jeonghoon Ha, Hongwoo Baek, Yang-Hao Chan, Fabian D. Natterer, Alline F. Myers, Joshua D. Schumacher, William G. Cullen, Albert V. Davydov, Young Kuk, M. Y. Chou, Nikolai B. Zhitenev, Joseph A. Stroscio

Abstract: We report a rectangular charge density wave (CDW) phase in strained 1T-VSe$_2$ thin films synthesized by molecular beam epitaxy on c-sapphire substrates. The observed CDW structure exhibits an unconventional rectangular 4a{\times}{\sqrt{3a}} periodicity, as opposed to the previously reported hexagonal $4a\times4a$ structure in bulk crystals and exfoliated thin layered samples. Tunneling spectrosco… ▽ More We report a rectangular charge density wave (CDW) phase in strained 1T-VSe$_2$ thin films synthesized by molecular beam epitaxy on c-sapphire substrates. The observed CDW structure exhibits an unconventional rectangular 4a{\times}{\sqrt{3a}} periodicity, as opposed to the previously reported hexagonal $4a\times4a$ structure in bulk crystals and exfoliated thin layered samples. Tunneling spectroscopy shows a strong modulation of the local density of states of the same $4a\times\sqrt{3}a$ CDW periodicity and an energy gap of $2Δ_{CDW}=(9.1\pm0.1)$ meV. The CDW energy gap evolves into a full gap at temperatures below 500 mK, indicating a transition to an insulating phase at ultra-low temperatures. First-principles calculations confirm the stability of both $4a\times4a$ and $4a\times\sqrt{3}a$ structures arising from soft modes in the phonon dispersion. The unconventional structure becomes preferred in the presence of strain, in agreement with experimental findings. △ Less

Submitted 20 July, 2017; originally announced July 2017.

Journal ref: Physical Review Materials, 1, 024005 (2017)

Surface Transport and Quantum Hall Effect in Ambipolar Black Phosphorus Double Quantum Wells

Authors: Son Tran, Jiawei Yang, Nathaniel Gillgren, Timothy Espiritu, Yanmeng Shi, Kenji Watanabe, Takashi Taniguchi, Seongphill Moon, Hongwoo Baek, Dmitry Smirnov, Marc Bockrath, Ruoyu Chen, Chun Ning Lau

Abstract: Quantum wells constitute one of the most important classes of devices in the study of 2D systems. In a double layer QW, the additional "which-layer" degree of freedom gives rise to celebrated phenomena such as Coulomb drag, Hall drag and exciton condensation. Here we demonstrate facile formation of wide QWs in few-layer black phosphorus devices that host double layers of charge carriers. In contra… ▽ More Quantum wells constitute one of the most important classes of devices in the study of 2D systems. In a double layer QW, the additional "which-layer" degree of freedom gives rise to celebrated phenomena such as Coulomb drag, Hall drag and exciton condensation. Here we demonstrate facile formation of wide QWs in few-layer black phosphorus devices that host double layers of charge carriers. In contrast to tradition QWs, each 2D layer is ambipolar, and can be tuned into n-doped, p-doped or intrinsic regimes. Fully spin-polarized quantum Hall states are observed on each layer, with enhanced Lande g-factor that is attributed to exchange interactions. Our work opens the door for using 2D semiconductors as ambipolar single, double or wide QWs with unusual properties such as high anisotropy. △ Less

Submitted 26 March, 2017; v1 submitted 14 March, 2017; originally announced March 2017.

Journal ref: Science Advances 3, e1603179 (2017)

Dualizable and semi-flat objects in abstract module categories

Authors: Rune Harder Bak

Abstract: In this paper, we define what it means for an object in an abstract module category to be dualizable and we give a homological description of the direct limit closure of the dualizable objects. Our description recovers existing results of Govorov and Lazard, Oberst and R{ö}hrl, and Christensen and Holm. When applied to differential graded modules over a differential graded algebra, our description… ▽ More In this paper, we define what it means for an object in an abstract module category to be dualizable and we give a homological description of the direct limit closure of the dualizable objects. Our description recovers existing results of Govorov and Lazard, Oberst and R{ö}hrl, and Christensen and Holm. When applied to differential graded modules over a differential graded algebra, our description yields that a DG-module is semi-flat if and only if it can be obtained as a direct limit of finitely generated semi-free DG-modules. We obtain similar results for graded modules over graded rings and for quasi-coherent sheaves over nice schemes. △ Less

Submitted 11 May, 2018; v1 submitted 9 July, 2016; originally announced July 2016.

Comments: 19 pages. Final version to appear in Math. Z. Title have been slightly modified. Major changes in exposition. References have been added. Typos have been corrected. Main theorem strengthened

MSC Class: 18E15 (Primary); 16E45; 18G35 (Secondary)

Efficient Hybrid White Organic Light-Emitting Diodes for Application of Triplet Harvesting with Simple Structure

Authors: Kyo Min Hwang, Song Eun Lee, Sungkyu Lee, Han Kyu Yoo, Hyun Jung Baek, Jwajin Kim, Seung Soo Yoon, Young Kwan Kim

Abstract: In this study, we fabricated hybrid white organic light-emitting diodes (WOLEDs) based on triplet harvesting with simple structure. All the hole transporting material and host in emitting layer (EML) of devices were utilized with same material by using N,N'-di-1-naphthalenyl-N,N'-diphenyl-[1,1':4',1":4",1"'-quaterphenyl]-4,4"'-diamine (4P-NPD) which were known to be blue fluorescent material. Simp… ▽ More In this study, we fabricated hybrid white organic light-emitting diodes (WOLEDs) based on triplet harvesting with simple structure. All the hole transporting material and host in emitting layer (EML) of devices were utilized with same material by using N,N'-di-1-naphthalenyl-N,N'-diphenyl-[1,1':4',1":4",1"'-quaterphenyl]-4,4"'-diamine (4P-NPD) which were known to be blue fluorescent material. Simple hybrid WOLEDs were fabricated three color with blue fluorescent and green, red phosphorescent materials. We was investigated the effect of triplet harvesting (TH) by exciton generation zone on simple hybrid WOLEDs. Characteristic of simple hybrid WOLEDs were dominant hole mobility, therefore exciton generation zone was expected in EML. Additionally, we was optimization thickness of hole transporting layer and electron transporting layer was fabricated a simple hybrid WOLEDs. Simple hybrid WOLED exhibits maximum luminous efficiency of 29.3 cd/A and maximum external quantum efficiency of 11.2%. Commission Internationale de l'Eclairage (International Commission on Illumination) coordinates of (0.45, 0.43) at about 10,000 cd/m^2. △ Less

Submitted 23 May, 2016; originally announced May 2016.

Comments: 16 page, 5 figures, 1 table

Scanning Tunneling Spectroscopy of Proximity Superconductivity in Epitaxial Multilayer Graphene

Authors: Fabian D. Natterer, Jeonghoon Ha, Hongwoo Baek, Duming Zhang, William G. Cullen, Nikolai B. Zhitenev, Young Kuk, Joseph A. Stroscio

Abstract: We report on spatial measurements of the superconducting proximity effect in epitaxial graphene induced by a graphene-superconductor interface. Superconducting aluminum films were grown on epitaxial multilayer graphene on SiC. The aluminum films were discontinuous with networks of trenches in the film morphology reaching down to exposed graphene terraces. Scanning tunneling spectra measured on the… ▽ More We report on spatial measurements of the superconducting proximity effect in epitaxial graphene induced by a graphene-superconductor interface. Superconducting aluminum films were grown on epitaxial multilayer graphene on SiC. The aluminum films were discontinuous with networks of trenches in the film morphology reaching down to exposed graphene terraces. Scanning tunneling spectra measured on the graphene terraces show a clear decay of the superconducting energy gap with increasing separation from the graphene-aluminum edges. The spectra were well described by Bardeen-Cooper-Schrieffer (BCS) theory. The decay length for the superconducting energy gap in graphene was determined to be greater than 400 nm. Deviations in the exponentially decaying energy gap were also observed on a much smaller length scale of tens of nanometers. △ Less

Submitted 22 February, 2016; originally announced February 2016.

Journal ref: Physical Review B 93, 045406 (2016)

Creating Nanostructured Superconductors On Demand by Local Current Annealing

Authors: Hongwoo Baek, Jeonghoon Ha, Duming Zhang, Bharath Natarajan, Jonathan P. Winterstein, Renu Sharma, Rongwei Hu, Kefeng Wang, Steven Ziemak, Johnpierre Paglione, Young Kuk, Nikolai B. Zhitenev, Joseph A. Stroscio

Abstract: Superconductivity results from a Bose condensate of Cooper-paired electrons with a macroscopic quantum wavefunction. Dramatic effects can occur when the region of the condensate is shaped and confined to the nanometer scale. Recent progress in nanostructured superconductors has revealed a route to topological superconductivity, with possible applications in quantum computing. However, challenges r… ▽ More Superconductivity results from a Bose condensate of Cooper-paired electrons with a macroscopic quantum wavefunction. Dramatic effects can occur when the region of the condensate is shaped and confined to the nanometer scale. Recent progress in nanostructured superconductors has revealed a route to topological superconductivity, with possible applications in quantum computing. However, challenges remain in controlling the shape and size of specific superconducting materials. Here, we report a new method to create nanostructured superconductors by partial crystallization of the half-Heusler material, YPtBi. Superconducting islands, with diameters in the range of 100 nm, were reproducibly created by local current annealing of disordered YPtBi in the tunneling junction of a scanning tunneling microscope (STM). We characterize the superconducting island properties by scanning tunneling spectroscopic measurements to determine the gap energy, critical temperature and field, coherence length, and vortex formations. These results show unique properties of a confined superconductor and demonstrate that this new method holds promise to create tailored superconductors for a wide variety of nanometer scale applications. △ Less

Submitted 1 October, 2015; originally announced October 2015.

Comments: 12 figures

Journal ref: Physical Review B 92, 094510 (2015)

External Modulation and Switching of Acoustic Phonons: Comparative Roles of Potential Distributions

Authors: H. Jeong, Y. D. Jho, S. H. Rhim, K. J. Yee, J. P. Shim, D. S. Lee, J. W. Ju, J. H. Baek, C. J. Stanton

Abstract: Acoustic phonons can be coherently generated by ultrafast displacive screening of potential gradients, often enhanced by the strong built-in piezoelectric fields, in wurtzite semiconductors. In such structures, transverse symmetry within the c plane hinders both the generation and detection of the transverse acoustic (TA) modes, and only longitudinal acoustic (LA) mode is generated. We show that e… ▽ More Acoustic phonons can be coherently generated by ultrafast displacive screening of potential gradients, often enhanced by the strong built-in piezoelectric fields, in wurtzite semiconductors. In such structures, transverse symmetry within the c plane hinders both the generation and detection of the transverse acoustic (TA) modes, and only longitudinal acoustic (LA) mode is generated. We show that even for c-GaN, the application of asymmetric potential distributions in the c plane can break the symmetry and selection rules, thus switching on the normally forbidden TA mode. This is in contrast to the LA mode, the strength of which varies with the symmetric potential distributions. By comparing transient differential reflectivity spectra in structures with and without asymmetric potential distributions, the role of the electrically attained anisotropy was further revealed by the digitized appearance of the TA mode, in clear contrast to the monotonic LA mode, and by modulations in the propagation velocities, optical birefringence, and geometrically varying sensitivities, the underlying mechanisms of which are modeled by electric-field-dependent perturbations of the dielectric tensors, incorporating the results of elastic modulations. △ Less

Submitted 25 June, 2015; originally announced June 2015.

A bijection between aperiodic palindromes and connected circulant graphs

Authors: Hunki Baek, Sejeong Bang, Dongseok Kim, Jaeun Lee

Abstract: In this paper, we show that there is a one-to-one correspondence between the set of compositions (resp. prime compositions) of $n$ and the set of circulant digraphs (resp. connected circulant digraphs) of order $n$. We also show that there is a one-to-one correspondence between the set of palindromes (resp. aperiodic palindromes) of $n$ and the set of circulant graphs (resp. connected circulant gr… ▽ More In this paper, we show that there is a one-to-one correspondence between the set of compositions (resp. prime compositions) of $n$ and the set of circulant digraphs (resp. connected circulant digraphs) of order $n$. We also show that there is a one-to-one correspondence between the set of palindromes (resp. aperiodic palindromes) of $n$ and the set of circulant graphs (resp. connected circulant graphs) of order $n$. As a corollary of this correspondence, we enumerate the number of connected circulant (di)graphs of order $n$. △ Less

Submitted 7 December, 2014; originally announced December 2014.

MSC Class: 05C30 05A17 05C25

Quasiparticle scattering from topological crystalline insulator SnTe (001) surface states

Authors: Duming Zhang, Hongwoo Baek, Jeonghoon Ha, Tong Zhang, Jonathan E. Wyrick, Albert V. Davydov, Young Kuk, Joseph A. Stroscio

Abstract: Recently, the topological classification of electronic states has been extended to a new class of matter known as topological crystalline insulators. Similar to topological insulators, topological crystalline insulators also have spin-momentum locked surface states; but they only exist on specific crystal planes that are protected by crystal reflection symmetry. Here, we report an ultra-low temper… ▽ More Recently, the topological classification of electronic states has been extended to a new class of matter known as topological crystalline insulators. Similar to topological insulators, topological crystalline insulators also have spin-momentum locked surface states; but they only exist on specific crystal planes that are protected by crystal reflection symmetry. Here, we report an ultra-low temperature scanning tunneling microscopy and spectroscopy study on topological crystalline insulator SnTe nanoplates grown by molecular beam epitaxy. We observed quasiparticle interference patterns on the SnTe (001) surface that can be interpreted in terms of electron scattering from the four Fermi pockets of the topological crystalline insulator surface states in the first surface Brillouin zone. A quantitative analysis of the energy dispersion of the quasiparticle interference intensity shows two high energy features related to the crossing point beyond the Lifshitz transition when the two neighboring low energy surface bands near the point merge. A comparison between the experimental and computed quasiparticle interference patterns reveals possible spin texture of the surface states. △ Less

Submitted 5 May, 2014; originally announced May 2014.

Journal ref: Physical Review B 89, 245445 (2014)

Temperature Evolution of Itinerant Ferromagnetism in SrRuO3 Probed by Optical Spectroscopy

Authors: D. W. Jeong, Hong Chul Choi, Choong H. Kim, Seo Hyoung Chang, C. H. Sohn, H. J. Park, T. D. Kang, Deok-Yong Cho, S. H. Baek, C. B. Eom, J. H. Shim, J. Yu, K. W. Kim, S. J. Moon, T. W. Noh

Abstract: The temperature ($T$) dependence of the optical conductivity spectra $σ(ω)$ of a single crystal SrRuO$_3$ thin film is studied over a $T$ range from 5 to 450 K. We observed significant $T$ dependence of the spectral weights of the charge transfer and interband $d$-$d$ transitions across the ferromagnetic Curie temperature ($T_c$ ~ 150 K). Such $T$ dependence was attributed to the increase in the R… ▽ More The temperature ($T$) dependence of the optical conductivity spectra $σ(ω)$ of a single crystal SrRuO$_3$ thin film is studied over a $T$ range from 5 to 450 K. We observed significant $T$ dependence of the spectral weights of the charge transfer and interband $d$-$d$ transitions across the ferromagnetic Curie temperature ($T_c$ ~ 150 K). Such $T$ dependence was attributed to the increase in the Ru spin moment, which is consistent with the results of density functional theory calculations. $T$ scans of $σ(Ω, T)$ at fixed frequencies $Ω$ reveal a clear $T^2$ dependence below $T_c$, demonstrating that the Stoner mechanism is involved in the evolution of the electronic structure. In addition, $σ(Ω, T)$ continues to evolve at temperatures above $T_c$, indicating that the local spin moment persists in the paramagnetic state. This suggests that SrRuO$_3$ is an intriguing oxide system with itinerant ferromagnetism. △ Less

Submitted 4 September, 2013; originally announced September 2013.

Journal ref: Physical Review Letters, 110, 247202 (2013)

Velocity estimation via registration-guided least-squares inversion

Authors: Hyoungsu Baek, Henri Calandra, Laurent Demanet

Abstract: This paper introduces an iterative scheme for acoustic model inversion where the notion of proximity of two traces is not the usual least-squares distance, but instead involves registration as in image processing. Observed data are matched to predicted waveforms via piecewise-polynomial warpings, obtained by solving a nonconvex optimization problem in a multiscale fashion from low to high frequenc… ▽ More This paper introduces an iterative scheme for acoustic model inversion where the notion of proximity of two traces is not the usual least-squares distance, but instead involves registration as in image processing. Observed data are matched to predicted waveforms via piecewise-polynomial warpings, obtained by solving a nonconvex optimization problem in a multiscale fashion from low to high frequencies. This multiscale process requires defining low-frequency augmented signals in order to seed the frequency sweep at zero frequency. Custom adjoint sources are then defined from the warped waveforms. The proposed velocity updates are obtained as the migration of these adjoint sources, and cannot be interpreted as the negative gradient of any given objective function. The new method, referred to as RGLS, is successfully applied to a few scenarios of model velocity estimation in the transmission setting. We show that the new method can converge to the correct model in situations where conventional least-squares inversion suffers from cycle-skipping and converges to a spurious model. △ Less

Submitted 18 April, 2013; originally announced April 2013.

Comments: 20 pages, 13 figures, 1 table

Conditioning bounds for traveltime tomography in layered media

Authors: Hyoungsu Baek, Laurent Demanet

Abstract: This paper revisits the problem of recovering a smooth, isotropic, layered wave speed profile from surface traveltime information. While it is classic knowledge that the diving (refracted) rays classically determine the wave speed in a weakly well-posed fashion via the Abel transform, we show in this paper that traveltimes of reflected rays do not contain enough information to recover the medium i… ▽ More This paper revisits the problem of recovering a smooth, isotropic, layered wave speed profile from surface traveltime information. While it is classic knowledge that the diving (refracted) rays classically determine the wave speed in a weakly well-posed fashion via the Abel transform, we show in this paper that traveltimes of reflected rays do not contain enough information to recover the medium in a well-posed manner, regardless of the discretization. The counterpart of the Abel transform in the case of reflected rays is a Fredholm kernel of the first kind which is shown to have singular values that decay at least root-exponentially. Kinematically equivalent media are characterized in terms of a sequence of matching moments. This severe conditioning issue comes on top of the well-known rearrangement ambiguity due to low velocity zones. Numerical experiments in an ideal scenario show that a waveform-based model inversion code fits data accurately while converging to the wrong wave speed profile. △ Less

Submitted 14 October, 2011; originally announced October 2011.

MSC Class: 45Q05

Polarity control of carrier injection at ferroelectric/metal interfaces for electrically switchable diode and photovoltaic effects

Authors: Daesu Lee, S. H. Baek, T. H. Kim, J. -G. Yoon, C. M. Folkman, C. B. Eom, T. W. Noh

Abstract: We investigated a switchable ferroelectric diode effect and its physical mechanism in Pt/BiFeO3/SrRuO3 thin-film capacitors. Our results of electrical measurements support that, near the Pt/BiFeO3 interface of as-grown samples, a defective layer (possibly, an oxygen-vacancy-rich layer) becomes formed and disturbs carrier injection. We therefore used an electrical training process to obtain ferroel… ▽ More We investigated a switchable ferroelectric diode effect and its physical mechanism in Pt/BiFeO3/SrRuO3 thin-film capacitors. Our results of electrical measurements support that, near the Pt/BiFeO3 interface of as-grown samples, a defective layer (possibly, an oxygen-vacancy-rich layer) becomes formed and disturbs carrier injection. We therefore used an electrical training process to obtain ferroelectric control of the diode polarity where, by changing the polarization direction using an external bias, we could switch the transport characteristics between forward and reverse diodes. Our system is characterized with a rectangular polarization hysteresis loop, with which we confirmed that the diode polarity switching occurred at the ferroelectric coercive voltage. Moreover, we observed a simultaneous switching of the diode polarity and the associated photovoltaic response dependent on the ferroelectric domain configurations. Our detailed study suggests that the polarization charge can affect the Schottky barrier at the ferroelectric/metal interfaces, resulting in a modulation of the interfacial carrier injection. The amount of polarization-modulated carrier injection can affect the transition voltage value at which a space-charge-limited bulk current-voltage (J-V) behavior is changed from Ohmic (i.e., J ~ V) to nonlinear (i.e., J ~ V^n with n \geq 2). This combination of bulk conduction and polarization-modulated carrier injection explains the detailed physical mechanism underlying the switchable diode effect in ferroelectric capacitors. △ Less

Submitted 16 August, 2011; originally announced August 2011.

Comments: Accepted for publication in Phys. Rev. B

Journal ref: Phys. Rev. B 84, 125305 (2011)