BZPEERReading, trust and collaboration for research
Menu

Discover

GraphFollow connections around a paper, topic or saved view.

Workspaces

HomePick up where your research flow left off.InboxHandle requests, replies and notifications from one place.CollectionsCurate reading lists, evidence sets and shareable dossiers.ResearchSee your private provenance graph, trail and imports.CollaborationMove from discovery into focused discussion rooms.PublishDraft, preview and publish full-text research articles.

Network

ExpertsFind specialists worth following or asking for help.CommunitiesJoin research circles organized around shared work.PartnersSee external organizations active around the same topics.

Opportunities

ListingsBrowse roles, calls and collaborations with clearer fit signals.

Account

Log inReturn to your reading and collaboration workspace.Create accountStart saving work, following people and joining teams.
Log inCreate account

Researcher profile

Jinwoo Hwang

Jinwoo Hwang contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate
cond-mat.mtrl-sciComputer Visioncond-mat.mes-hallcond-mat.dis-nnCryptography and Securityeess.SYEmerging TechnologiesMachine Learningphysics.app-phSystems and Control

Trust snapshot

Quick read

Trust 21 - EmergingVerification L1Unclaimed author
11works
0followers
10topics
4close collaborators

Actions

Decide how to stay connected

Follow researcher0

Identity and collaboration

How to connect with this researcher

Claiming links this public author record to a researcher profile and unlocks direct collaboration workflows.

Log in to claim

Direct collaboration

Open a focused conversation when the fit is right

Claim this author entity first to unlock direct invitations.

Research graph

See the researcher in context

Open full explorer

Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.

Building this graph slice

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Published work

11 published item(s)

preprint2026arXiv

Understanding the Performance Behaviors of End-to-End Protein Design Pipelines on GPUs

Recent computational advances enable protein design pipelines to run end-to-end on GPUs, yet their heterogeneous computational behaviors remain undercharacterized at the system level. We implement and profile a representative pipeline at both component and full-pipeline granularities across varying inputs and hyperparameters. Our characterization identifies generally low GPU utilization and high sensitivity to sequence length and sampling strategies. We outline future research directions based on these insights and release an open-source pipeline and profiling scripts to facilitate further studies.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Anisotropic Magnetoresistance and Nontrivial Spin Hall Magnetoresistance in Pt/$α$-Fe$_2$O$_3$ Bilayers

To date, magnetic proximity effect (MPE) has only been conclusively observed in ferromagnet (FM) based systems. We report the observation of anomalous Hall effect and anisotropic magnetoresistance in angular dependent magnetoresistance (ADMR) measurements in Pt on antiferromagnetic (AF) $α$-Fe$_2$O$_3$(0001) epitaxial films at 10 K, which provide evidence for the MPE. The Néel order of $α$-Fe$_2$O$_3$ and the induced magnetization in Pt show a unique ADMR compared with all other FM and AF systems. A macrospin response model is established and can explain the AF spin configuration and all main ADMR features in the Pt/$α$-Fe$_2$O$_3$ bilayers.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Booster-SHOT: Boosting Stacked Homography Transformations for Multiview Pedestrian Detection with Attention

Improving multi-view aggregation is integral for multi-view pedestrian detection, which aims to obtain a bird's-eye-view pedestrian occupancy map from images captured through a set of calibrated cameras. Inspired by the success of attention modules for deep neural networks, we first propose a Homography Attention Module (HAM) which is shown to boost the performance of existing end-to-end multiview detection approaches by utilizing a novel channel gate and spatial gate. Additionally, we propose Booster-SHOT, an end-to-end convolutional approach to multiview pedestrian detection incorporating our proposed HAM as well as elements from previous approaches such as view-coherent augmentation or stacked homography transformations. Booster-SHOT achieves 92.9% and 94.2% for MODA on Wildtrack and MultiviewX respectively, outperforming the state-of-the-art by 1.4% on Wildtrack and 0.5% on MultiviewX, achieving state-of-the-art performance overall for standard evaluation metrics used in multi-view pedestrian detection.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

CoVA: Exploiting Compressed-Domain Analysis to Accelerate Video Analytics

Modern retrospective analytics systems leverage cascade architecture to mitigate bottleneck for computing deep neural networks (DNNs). However, the existing cascades suffer two limitations: (1) decoding bottleneck is either neglected or circumvented, paying significant compute and storage cost for pre-processing; and (2) the systems are specialized for temporal queries and lack spatial query support. This paper presents CoVA, a novel cascade architecture that splits the cascade computation between compressed domain and pixel domain to address the decoding bottleneck, supporting both temporal and spatial queries. CoVA cascades analysis into three major stages where the first two stages are performed in compressed domain while the last one in pixel domain. First, CoVA detects occurrences of moving objects (called blobs) over a set of compressed frames (called tracks). Then, using the track results, CoVA prudently selects a minimal set of frames to obtain the label information and only decode them to compute the full DNNs, alleviating the decoding bottleneck. Lastly, CoVA associates tracks with labels to produce the final analysis results on which users can process both temporal and spatial queries. Our experiments demonstrate that CoVA offers 4.8x throughput improvement over modern cascade systems, while imposing modest accuracy loss.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Electrical Switching of Tristate Antiferromagnetic Néel Order in $α$-Fe$_{2}$O$_{3}$ Epitaxial Films

The ability to manipulate antiferromagnetic (AF) moments is a key requirement for the emerging field of antiferromagnetic spintronics. Electrical switching of bi-state AF moments has been demonstrated in metallic AFs, CuMnAs and Mn$_2$Au. Recently, current-induced "saw-tooth" shaped Hall resistance was reported in Pt/NiO bilayers, while its mechanism is under debate. Here, we report the first demonstration of convincing, non-decaying, step-like electrical switching of tri-state Néel order in Pt/$α$-Fe$_2$O$_3$ bilayers. Our experimental data, together with Monte-Carlo simulations, reveal the clear mechanism of the switching behavior of $α$-Fe$_2$O$_3$ Néel order among three stable states. We also show that the observed "saw-tooth" Hall resistance is due to an artifact of Pt, not AF switching, while the signature of AF switching is step-like Hall signals. This demonstration of electrical control of magnetic moments in AF insulator (AFI) films will greatly expand the scope of AF spintronics by leveraging the large family of AFIs.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Kinetically-controlled epitaxial growth of Fe$_3$GeTe$_2$ van der Waals ferromagnetic films

We demonstrate that kinetics play an important role in the epitaxial growth of Fe$_3$GeTe$_2$ (FGT) van der Waals (vdW) ferromagnetic films by molecular beam epitaxy. By varying the deposition rate, we control the formation or suppression of an initial tellurium-deficient non-van der Waals phase (Fe$_3$Ge$_2$) prior to realizing epitaxial growth of the vdW FGT phase. Using cross-sectional scanning transmission electron microscopy and scanning tunneling microscopy, we optimize the FGT films to have atomically smooth surfaces and abrupt interfaces with the Ge(111) substrate. The magnetic properties of our high quality material are confirmed through magneto-optic, magnetotransport, and spin-polarized STM studies. Importantly, this demonstrates how the interplay of energetics and kinetics can help tune the re-evaporation rate of chalcogen atoms and interdiffusion from the underlayer, which paves the way for future studies of van der Waals epitaxy.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Multiplication of freestanding semiconductor membranes from a single wafer by advanced remote epitaxy

Freestanding single-crystalline membranes are an important building block for functional electronics. Especially, compounds semiconductor membranes such as III-N and III-V offer great opportunities for optoelectronics, high-power electronics, and high-speed computing. Despite huge efforts to produce such membranes by detaching epitaxial layers from donor wafers, however, it is still challenging to harvest epitaxial layers using practical processes. Here, we demonstrate a method to grow and harvest multiple epitaxial membranes with extremely high throughput at the wafer scale. For this, 2D materials are directly formed on III-N and III-V substrates in epitaxy systems, which enables an advanced remote epitaxy scheme comprised of multiple alternating layers of 2D materials and epitaxial layers that can be formed by a single epitaxy run. Each epilayer in the multi-stack structure is then harvested by layer-by-layer peeling, producing multiple freestanding membranes with unprecedented throughput from a single wafer. Because 2D materials allow peeling at the interface without damaging the epilayer or the substrate, wafers can be reused for subsequent membrane production. Therefore, this work represents a meaningful step toward high-throughput and low-cost production of single-crystal membranes that can be heterointegrated.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Privacy Safe Representation Learning via Frequency Filtering Encoder

Deep learning models are increasingly deployed in real-world applications. These models are often deployed on the server-side and receive user data in an information-rich representation to solve a specific task, such as image classification. Since images can contain sensitive information, which users might not be willing to share, privacy protection becomes increasingly important. Adversarial Representation Learning (ARL) is a common approach to train an encoder that runs on the client-side and obfuscates an image. It is assumed, that the obfuscated image can safely be transmitted and used for the task on the server without privacy concerns. However, in this work, we find that training a reconstruction attacker can successfully recover the original image of existing ARL methods. To this end, we introduce a novel ARL method enhanced through low-pass filtering, limiting the available information amount to be encoded in the frequency domain. Our experimental results reveal that our approach withstands reconstruction attacks while outperforming previous state-of-the-art methods regarding the privacy-utility trade-off. We further conduct a user study to qualitatively assess our defense of the reconstruction attack.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Experimental determination of the valence band offsets of $ZnGeN_2$ and $ZnGe_{0.94}Ga_{0.12}N_2$ with $GaN$

A predicted type-II staggered band alignment with an approximately $1.4 eV$ valence band offset at the $ZnGeN_2/GaN$ heterointerface has inspired novel band-engineered $III-N/ZnGeN_2$ heterostructure-based device designs for applications in high performance optoelectronics. We report on the determination of the valence band offset between metalorganic chemical vapor deposition grown $(ZnGe)_{1-x}Ga_{2x}N_2$, for $x = 0$ and $0.06$, and $GaN$ using X-ray photoemission spectroscopy. The valence band of $ZnGeN_2$ was found to lie $1.45-1.65 eV$ above that of $GaN$. This result agrees well with the value predicted by first-principles density functional theory calculations using the local density approximation for the potential profile and quasiparticle self-consistent GW calculations of the band edge states relative to the potential. For $(ZnGe)_{0.94}Ga_{0.12}N_2$ the value was determined to be $1.29 eV$, $~10-20\%$ lower than that of $ZnGeN_2$. The experimental determination of the large band offset between $ZnGeN_2$ and $GaN$ provides promising alternative solutions to address challenges faced with pure III-nitride-based structures and devices.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Coherent Growth and Characterization of van der Waals 1T-VSe$_2$ Layers on GaAs(111)B Using Molecular Beam Epitaxy

We report epitaxial growth of vanadium diselenide (VSe$_2$) thin films in the octahedrally-coordinated (1T) structure on GaAs(111)B substrates by molecular beam epitaxy. Film thickness from a single monolayer (ML) up to 30 ML is demonstrated. Structural and chemical studies using by x-ray diffraction, transmission electron microscopy, scanning tunneling microscopy and x-ray photoelectron spectroscopy indicate high quality thin films. Further studies show that monolayer VSe$_2$ films on GaAs are not air-stable and are susceptible to oxidation within a matter of hours, which indicates that a protective capping layer should be employed for device applications. This work demonstrates that VSe$_2$, a candidate van der Waals material for possible spintronic and electronic applications, can be integrated with III-V semiconductors via epitaxial growth for 2D/3D hybrid devices.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Structural Heterogeneity, Ductility, and Glass Forming Ability of Zr-Based Metallic Glasses

We show the correlation between nanoscale structural heterogeneity and mechanical property and glass forming ability of Zr-based metallic glasses (MGs). Detailed parameters of medium range ordering (MRO) that constitutes the structural heterogeneity, including the type, size, and volume fraction of MRO domains determined using 4-dimensional scanning transmission electron microscopy, directly correlate with the ductility and glass forming ability of Zr-Cu-Co-Al MGs. Mesoscale deformation simulation incorporating the experimentally determined MRO confirms that the diverse types and sizes of MRO can significantly influence the MGs' mechanical behavior.

0 citations0 reviewsNo code linkedOpen access