arXiv subjects

Youngwook Kim

Publications and source records attributed to Youngwook Kim.

At least 37 records · Page 2Linked to original sources

arXiv2023

Orbitally controlled quantum Hall states in decoupled two-bilayer graphene sheets

We report on integer and fractional quantum Hall states in a stack of two twisted Bernal bilayer graphene sheets. By exploiting the momentum mismatch in reciprocal space, we suppress single particle tunneling between both bilayers. Since the bilayers are spatially separated by only 0.34 nm, the stack benefits from strong interlayer Coulomb interactions. These interactions can cause the formation of a Bose-Einstein condensate. Indeed, such a condensate is observed for half filling in each bilayer sheet. However, only when the partially filled levels have orbital index 1. It is absent for partially filled levels with orbital index 0. This discrepancy is tentatively attributed to the role of skyrmion/anti-skyrmion pair excitations and the dependence of the energy of these excitations on the orbital index. The application of asymmetric top and bottom gate voltages enables to influence the orbital nature of the electronic states of the graphene bilayers at the chemical potential and to navigate in an orbital mixed space. The latter hosts an even denominator fractional quantum Hall state at total filling -3/2. Our observations suggest a unique edge reconstruction involving both electrons and chiral p-wave composite fermions.

cond-mat.mes-hall↗

arXiv2023

Bridging the Gap between Model Explanations in Partially Annotated Multi-label Classification

Due to the expensive costs of collecting labels in multi-label classification datasets, partially annotated multi-label classification has become an emerging field in computer vision. One baseline approach to this task is to assume unobserved labels as negative labels, but this assumption induces label noise as a form of false negative. To understand the negative impact caused by false negative labels, we study how these labels affect the model's explanation. We observe that the explanation of two models, trained with full and partial labels each, highlights similar regions but with different scaling, where the latter tends to have lower attribution scores. Based on these findings, we propose to boost the attribution scores of the model trained with partial labels to make its explanation resemble that of the model trained with full labels. Even with the conceptually simple approach, the multi-label classification performance improves by a large margin in three different datasets on a single positive label setting and one on a large-scale partial label setting. Code is available at https://github.com/youngwk/BridgeGapExplanationPAMC.

cs.CV↗

arXiv2022

Robust Interlayer-Coherent Quantum Hall States in Twisted Bilayer Graphene

We introduce a novel two-dimensional electronic system with ultrastrong interlayer interactions, namely twisted bilayer graphene with a large twist angle, as an ideal ground for realizing interlayer-coherent excitonic condensates. In these systems, subnanometer atomic separation between the layers allows significant interlayer interactions, while interlayer electron tunneling is geometrically suppressed due to the large twist angle. By fully exploiting these two features we demonstrate that a sequence of odd-integer quantum Hall states with interlayer coherence appears at the second Landau level (N = 1). Notably the energy gaps for these states are of order 1 K, which is several orders of magnitude greater than those in GaAs. Furthermore, a variety of quantum Hall phase transitions are observed experimentally. All the experimental observations are largely consistent with our phenomenological model calculations. Hence, we establish that a large twist angle system is an excellent platform for high-temperature excitonic condensation.

cond-mat.mes-hall↗

arXiv2022

Modularized Transfer Learning with Multiple Knowledge Graphs for Zero-shot Commonsense Reasoning

Commonsense reasoning systems should be able to generalize to diverse reasoning cases. However, most state-of-the-art approaches depend on expensive data annotations and overfit to a specific benchmark without learning how to perform general semantic reasoning. To overcome these drawbacks, zero-shot QA systems have shown promise as a robust learning scheme by transforming a commonsense knowledge graph (KG) into synthetic QA-form samples for model training. Considering the increasing type of different commonsense KGs, this paper aims to extend the zero-shot transfer learning scenario into multiple-source settings, where different KGs can be utilized synergetically. Towards this goal, we propose to mitigate the loss of knowledge from the interference among the different knowledge sources, by developing a modular variant of the knowledge aggregation as a new zero-shot commonsense reasoning framework. Results on five commonsense reasoning benchmarks demonstrate the efficacy of our framework, improving the performance with multiple KGs.

cs.AI↗

arXiv2022

Preparation of large Cu3Sn single crystal by Czochralski method

Cu3Sn was recently predicted to host topological Dirac fermions, but related research is still in its infancy. The growth of large and high-quality Cu3Sn single crystals is, therefore, highly desired to investigate the possible topological properties. In this work, we report the single crystal growth of Cu3Sn by Czochralski (CZ) method. Crystal structure, chemical composition, and transport properties of Cu3Sn single crystals were analyzed to verify the crystal quality. Notably, compared to the mm-sized crystals from a molten Sn-flux, the cm-sized crystals obtained by the CZ method are free from contamination from flux materials, paving the way for the follow-up works.

cond-mat.mtrl-sci↗

arXiv2022

Large Loss Matters in Weakly Supervised Multi-Label Classification

Weakly supervised multi-label classification (WSML) task, which is to learn a multi-label classification using partially observed labels per image, is becoming increasingly important due to its huge annotation cost. In this work, we first regard unobserved labels as negative labels, casting the WSML task into noisy multi-label classification. From this point of view, we empirically observe that memorization effect, which was first discovered in a noisy multi-class setting, also occurs in a multi-label setting. That is, the model first learns the representation of clean labels, and then starts memorizing noisy labels. Based on this finding, we propose novel methods for WSML which reject or correct the large loss samples to prevent model from memorizing the noisy label. Without heavy and complex components, our proposed methods outperform previous state-of-the-art WSML methods on several partial label settings including Pascal VOC 2012, MS COCO, NUSWIDE, CUB, and OpenImages V3 datasets. Various analysis also show that our methodology actually works well, validating that treating large loss properly matters in a weakly supervised multi-label classification. Our code is available at https://github.com/snucml/LargeLossMatters.

cs.CV↗

arXiv2022

Growth of delafossite CuAlO2 single crystals in a reactive crucible

Delafossite oxide CuAlO2 has engaged great attention as a promising p-type conducting oxide. In this work, high-quality CuAlO2 single crystals with a size of several millimeters (mm) are successfully achieved with a reactive crucible melting method. The crystals are characterized by X-ray diffraction, scanning electron microscopy with energy-dispersive spectroscopy, transport measurement, and magnetic susceptibility measurement. The grown single crystals are free of contamination from a copper oxide flux. This work provides a new approach to growing high-quality delafossite oxide CuAlO2 with a few mm size.

cond-mat.mtrl-sci↗

arXiv2022

Dual Task Framework for Improving Persona-grounded Dialogue Dataset

This paper introduces a simple yet effective data-centric approach for the task of improving persona-conditioned dialogue agents. Prior model-centric approaches unquestioningly depend on the raw crowdsourced benchmark datasets such as Persona-Chat. In contrast, we aim to fix annotation artifacts in benchmarking, which is orthogonally applicable to any dialogue model. Specifically, we augment relevant personas to improve dialogue dataset/agent, by leveraging the primal-dual structure of the two tasks, predicting dialogue responses and personas based on each other. Experiments on Persona-Chat show that our approach outperforms pre-trained LMs by an 11.7 point gain in terms of accuracy.

cs.CL↗

arXiv2022

TrustAL: Trustworthy Active Learning using Knowledge Distillation

Active learning can be defined as iterations of data labeling, model training, and data acquisition, until sufficient labels are acquired. A traditional view of data acquisition is that, through iterations, knowledge from human labels and models is implicitly distilled to monotonically increase the accuracy and label consistency. Under this assumption, the most recently trained model is a good surrogate for the current labeled data, from which data acquisition is requested based on uncertainty/diversity. Our contribution is debunking this myth and proposing a new objective for distillation. First, we found example forgetting, which indicates the loss of knowledge learned across iterations. Second, for this reason, the last model is no longer the best teacher -- For mitigating such forgotten knowledge, we select one of its predecessor models as a teacher, by our proposed notion of "consistency". We show that this novel distillation is distinctive in the following three aspects; First, consistency ensures to avoid forgetting labels. Second, consistency improves both uncertainty/diversity of labeled data. Lastly, consistency redeems defective labels produced by human annotators.

cs.LG↗

arXiv2021

Odd integer quantum Hall states with interlayer coherence in twisted bilayer graphene

We report on the quantum Hall effect in two stacked graphene layers rotated by 2 degree. The tunneling strength among the layers can be varied from very weak to strong via the mechanism of magnetic breakdown when tuning the density. Odd-integer quantum Hall physics is not anticipated in the regime of suppressed tunneling for balanced layer densities, yet it is observed. We interpret this as a signature of Coulomb interaction induced interlayer coherence and Bose Einstein condensation of excitons that form at half filling of each layer. A density imbalance gives rise to reentrant behavior due to a phase transition from the interlayer coherent state to incompressible behavior caused by simultaneous condensation of both layers in different quantum Hall states. With increasing overall density, magnetic breakdown gains the upper hand. As a consequence of the enhanced interlayer tunneling, the interlayer coherent state and the phase transition vanish.

cond-mat.mes-hall↗

arXiv2020

Rectification by hydrodynamic flow in an encapsulated graphene Tesla valve

Systems in which interparticle interactions prevail can be described by hydrodynamics. This regime is typically difficult to access in the solid state for electrons. However, the high purity of encapsulated graphene combined with its advantageous phonon properties make it possible, and hydrodynamic corrections to the conductivity of graphene have been observed. Examples include electron whirlpools, enhanced flow through constrictions as well as a Poiseuille flow profile. An electronic device relying specifically on viscous behaviour and acting as a viscometer has however been lacking. Here, we implement the analogue of the Tesla valve. It exhibits nonreciprocal transport and can be regarded as an electronic viscous diode. Rectification occurs at carrier densities and temperatures consistent with the hydrodynamic regime, and disappears both in the ballistic and diffusive transport regimes. In a device in which the electrons are exposed to a Moiré superlattice, the Lifshitz transition when crossing the Van Hove singularity is observed in the rectifying behaviour.

cond-mat.mes-hall↗

arXiv2019

Reliable postprocessing improvement of van der Waals heterostructures

The successful assembly of heterostructures consisting of several layers of different 2D materials in arbitrary order by exploiting van der Waals forces has truly been a game changer in the field of low dimensional physics. For instance, the encapsulation of graphene or MoS2 between atomically flat hexagonal boron nitride (hBN) layers with strong affinity and graphitic gates that screen charge impurity disorder provided access to a plethora of interesting physical phenomena by drastically boosting the device quality. The encapsulation is accompanied by a self-cleansing effect at the interfaces. The otherwise predominant charged impurity disorder is minimized and random strain fluctuations ultimately constitute the main source of residual disorder. Despite these advances, the fabricated heterostructures still vary notably in their performance. While some achieve record mobilities, others only possess mediocre quality. Here, we report a reliable method to improve fully completed van der Waals heterostructure devices with a straightforward post-processing surface treatment based on thermal annealing and contact mode AFM. The impact is demonstrated by comparing magnetotransport measurements before and after the AFM treatment on one and the same device as well as on a larger set of treated and untreated devices to collect device statistics. Both the low temperature properties as well as the room temperature electrical characteristics, as relevant for applications, improve on average substantially. We surmise that the main beneficial effect arises from reducing nanometer scale corrugations at the interfaces, i.e. the detrimental impact of random strain fluctuations.

cond-mat.mtrl-sci↗

arXiv2019

Even denominator fractional quantum Hall states in higher Landau levels of graphene

An important development in the field of the fractional quantum Hall effect has been the proposal that the 5/2 state observed in the Landau level with orbital index $n = 1$ of two dimensional electrons in a GaAs quantum well originates from a chiral $p$-wave paired state of composite fermions which are topological bound states of electrons and quantized vortices. This state is theoretically described by a "Pfaffian" wave function or its hole partner called the anti-Pfaffian, whose excitations are neither fermions nor bosons but Majorana quasiparticles obeying non-Abelian braid statistics. This has inspired ideas on fault-tolerant topological quantum computation and has also instigated a search for other states with exotic quasiparticles. Here we report experiments on monolayer graphene that show clear evidence for unexpected even-denominator fractional quantum Hall physics in the $n=3$ Landau level. We numerically investigate the known candidate states for the even-denominator fractional quantum Hall effect, including the Pfaffian, the particle-hole symmetric Pfaffian, and the 221-parton states, and conclude that, among these, the 221-parton appears a potentially suitable candidate to describe the experimentally observed state. Like the Pfaffian, this state is believed to harbour quasi-particles with non-Abelian braid statistics

cond-mat.mes-hall↗

arXiv2019

Nanoscale imaging of equilibrium quantum Hall edge currents and of the magnetic monopole response in graphene

The recently predicted topological magnetoelectric effect and the response to an electric charge that mimics an induced mirror magnetic monopole are fundamental attributes of topological states of matter with broken time reversal symmetry. Using a SQUID-on-tip, acting simultaneously as a tunable scanning electric charge and as ultrasensitive nanoscale magnetometer, we induce and directly image the microscopic currents generating the magnetic monopole response in a graphene quantum Hall electron system. We find a rich and complex nonlinear behavior governed by coexistence of topological and nontopological equilibrium currents that is not captured by the monopole models. Furthermore, by utilizing a tuning fork that induces nanoscale vibrations of the SQUID-on-tip, we directly image the equilibrium currents of individual quantum Hall edge states for the first time. We reveal that the edge states that are commonly assumed to carry only a chiral downstream current, in fact carry a pair of counterpropagating currents, in which the topological downstream current in the incompressible region is always counterbalanced by heretofore unobserved nontopological upstream current flowing in the adjacent compressible region. The intricate patterns of the counterpropagating equilibrium-state orbital currents provide new insights into the microscopic origins of the topological and nontopological charge and energy flow in quantum Hall systems.

cond-mat.mes-hall↗

arXiv2019

Spin-split band hybridization in graphene proximitized with $α$-RuCl$_3$ nanosheets

Proximity effects induced in the 2D Dirac material graphene potentially open access to novel and intriguing physical phenomena. Thus far, the coupling between graphene and ferromagnetic insulators has been experimentally established. However, only very little is known about graphene's interaction with antiferromagnetic insulators. Here, we report a low temperature study of the electronic properties of high quality van der Waals heterostructures composed of a single graphene layer proximitized with $α$-RuCl$_3$. The latter is known to become antiferromagnetically ordered below 10 K. Shubnikov de Haas oscillations in the longitudinal resistance together with Hall resistance measurements provide clear evidence for a band realignment that is accompanied by a transfer of electrons originally occupying the graphene's spin degenerate Dirac cones into $α$-RuCl$_3$ band states with in-plane spin polarization. Left behind are holes in two separate Fermi pockets, only the dispersion of one of which is distorted near the Fermi energy due to spin selective hybridization with these spin polarized $α$-RuCl$_3$ band states. This interpretation is supported by our DFT calculations. An unexpected damping of the quantum oscillations as well as a zero field resistance upturn close to the N$é$el temperature of $α$-RuCl$_3$ suggests the onset of additional spin scattering due to spin fluctuations in the $α$-RuCl$_3$.

cond-mat.mtrl-sci↗

arXiv2019

Single spin resonance in a van der Waals embedded paramagnetic defect

Spins constitute a group of quantum objects forming a key resource in modern quantum technology. Two-dimensional (2D) van der Waals materials are of fundamental interest for studying nanoscale magnetic phenomena. However, isolating singular paramagnetic spins in 2D systems is challenging. We report here on a quantum emitting source embedded within hexgonal boron nitride (h-BN) exhibiting optical magnetic resonance (ODMR). We extract an isotropic $g$ factor close to 2 and derive an upper bound for a zero field splitting (ZFS) ($\leq$ 4 MHz). Photoluminescence (PL) behavior under temperature cycling using different excitations is presented, assigning probable zero phonon lines (ZPLs) / phonon side band (PSBs) to emission peaks, compatible with h-BN's phonon density of states, indicating their intrinsic nature. Narrow and inhomogeneous broadened ODMR lines differ significantly from monoatomic vacancy defect lines known in literature. We derive a hyperfine coupling of around 10 MHz. Its angular dependence indicates an unpaired electron in an out-of-plane $π$-orbital, probably originating from an additional substitutional carbon impurity or other low mass atom. We determine the spin relaxation time $T_1$ to be around 17 $μ$s.

cond-mat.mes-hall↗

arXiv2017

Quantum light in curved low dimensional hexagonal boron nitride systems

Low-dimensional wide bandgap semiconductors open a new playing field in quantum optics using sub-bandgap excitation. In this field, hexagonal boron nitride (h-BN) has been reported to host single quantum emitters (QEs), linking QE density to perimeters. Furthermore, curvature/perimeters in transition metal dichalcogenides (TMDCs) have demonstrated a key role in QE formation. We investigate a curvature-abundant BN system - quasi one-dimensional BN nanotubes (BNNTs) fabricated via a catalyst-free method. We find that non-treated BNNT is an abundant source of stable QEs and analyze their emission features down to single nanotubes, comparing dispersed/suspended material. Combining high spatial resolution of a scanning electron microscope, we categorize and pin-point emission origin to a scale of less than 20 nm, giving us a one-to-one validation of emission source with dimensions smaller than the laser excitation wavelength, elucidating nano-antenna effects. Two emission origins emerge: hybrid/entwined BNNT. By artificially curving h-BN flakes, similar QE spectral features are observed. The impact on emission of solvents used in commercial products and curved regions is also demonstrated. The 'out of the box' availability of QEs in BNNT, lacking processing contamination, is a milestone for unraveling their atomic features. These findings open possibilities for precision engineering of QEs, puts h-BN under a similar 'umbrella' of TMDC's QEs and provides a model explaining QEs spatial localization/formation using electron/ion irradiation and chemical etching.

cond-mat.mes-hall↗

arXiv2016

Broken-Symmetry Quantum Hall States in Twisted Bilayer Graphene

Twisted bilayer graphene offers a unique bilayer two-dimensional-electron system where the layer separation is only in sub-nanometer scale. Unlike Bernal-stacked bilayer, the layer degree of freedom is disentangled from spin and valley, providing eight-fold degeneracy in the low energy states. We have investigated broken-symmetry quantum Hall (QH) states and their transitions due to the interplay of the relative strength of valley, spin and layer polarizations in twisted bilayer graphene. The energy gaps of the broken-symmetry QH states show an electron-hole asymmetric behaviour, and their dependence on the induced displacement field are opposite between even and odd filling factor states. These results strongly suggest that the QH states with broken valley and spin symmetries for individual layer become hybridized via interlayer tunnelling, and the hierarchy of the QH states is sensitive to both magnetic field and displacement field due to charge imbalance between layers.

cond-mat.mes-hall↗