arXiv subjects

Jungeun Kim

Publications and source records attributed to Jungeun Kim.

15 recordsLinked to original sources

arXiv2026

Efficient Computation of Maximum Flexi-Clique in Networks

Discovering large cohesive subgraphs is a key task for graph mining. Existing models, such as clique, k-plex, and {\gamma}-quasi-clique, use fixed density thresholds that overlook the natural decay of connectivity as the subgraph size increases. The Flexi-clique model overcomes this limitation by imposing a degree constraint that grows sub-linearly with subgraph size. We provide the algorithmic study of Flexi-clique, proving its NP-hardness and analysing its non-hereditary properties. To address its computational challenge, we propose the Flexi-Prune Algorithm FPA, a fast heuristic using core-based seeding and connectivity-aware pruning, and the Efficient Branch-and-Bound Algorithm EBA, an exact framework enhanced with multiple pruning rules. Experiments on large real-world and synthetic networks demonstrate that FPA achieves near-optimal quality at much lower cost, while EBA efficiently computes exact solutions. Flexi-clique thus provides a practical and scalable model for discovering large, meaningful subgraphs in complex networks.

cs.SI↗

arXiv2025

Experimental Analysis and Evaluation of Cohesive Subgraph Discovery

Retrieving cohesive subgraphs in networks is a fundamental problem in social network analysis and graph data management. These subgraphs can be used for marketing strategies or recommendation systems. Despite the introduction of numerous models over the years, a systematic comparison of their performance, especially across varied network configurations, remains unexplored. In this study, we evaluated various cohesive subgraph models using task-based evaluations and conducted extensive experimental studies on both synthetic and real-world networks. Thus, we unveil the characteristics of cohesive subgraph models, highlighting their efficiency and applicability. Our findings not only provide a detailed evaluation of current models but also lay the groundwork for future research by shedding light on the balance between the interpretability and cohesion of the subgraphs. This research guides the selection of suitable models for specific analytical needs and applications, providing valuable insights.

cs.SI↗

arXiv2025

Cohesive Subgraph Discovery in Hypergraphs: A Locality-Driven Indexing Framework

Hypergraphs, increasingly utilised for modelling complex and diverse relationships in modern networks, gain much attention representing intricate higher-order interactions. Among various challenges, cohesive subgraph discovery is one of the fundamental problems and offers deep insights into these structures, yet the task of selecting appropriate parameters is an open question. To handle that question, we aim to design an efficient indexing structure to retrieve cohesive subgraphs in an online manner. The main idea is to enable the discovery of corresponding structures within a reasonable time without the need for exhaustive graph traversals. This work can facilitate efficient and informed decision-making in diverse applications based on a comprehensive understanding of the entire network landscape. Through extensive experiments on real-world networks, we demonstrate the superiority of our proposed indexing technique.

cs.SI↗

arXiv2025

High-Frequency Power Spectrum of AGN NGC 4051 Revealed by NICER

Variability studies offer a compelling glimpse into black hole dynamics, and NICER's (Neutron Star Interior Composition Explorer) remarkable temporal resolution propels us even further. NICER observations of an Active Galactic Nucleus (AGN), NGC 4051, have charted the geometry of the emission region of the central supermassive black hole. Our investigation of X-ray variability in NGC 4051 has detected extreme variations spanning a factor of 40 to 50 over a mere 10 to 12 hours. For the first time, we have constrained the X-ray Power Spectral Density (PSD) of the source to 0.1 Hz, corresponding to a temporal frequency of 10,000 Hz in a galactic X-ray binary (GXRB) with a mass of 10 M_{\odot}. No extra high-frequency break/bend or any quasi-periodic oscillations are found. Through detailed analysis of energy-dependent PSDs, we found that the PSD normalization, the high-frequency PSD slope as well as the bending frequency remains consistent across all energies within the 0.3-3 keV band, revealing the presence of a constant temperature corona. These significant findings impose critical constraints on current models of X-ray emission and variability in AGN.

astro-ph.HE↗

arXiv2024

DiffSLT: Enhancing Diversity in Sign Language Translation via Diffusion Model

Sign language translation (SLT) is challenging, as it involves converting sign language videos into natural language. Previous studies have prioritized accuracy over diversity. However, diversity is crucial for handling lexical and syntactic ambiguities in machine translation, suggesting it could similarly benefit SLT. In this work, we propose DiffSLT, a novel gloss-free SLT framework that leverages a diffusion model, enabling diverse translations while preserving sign language semantics. DiffSLT transforms random noise into the target latent representation, conditioned on the visual features of input video. To enhance visual conditioning, we design Guidance Fusion Module, which fully utilizes the multi-level spatiotemporal information of the visual features. We also introduce DiffSLT-P, a DiffSLT variant that conditions on pseudo-glosses and visual features, providing key textual guidance and reducing the modality gap. As a result, DiffSLT and DiffSLT-P significantly improve diversity over previous gloss-free SLT methods and achieve state-of-the-art performance on two SLT datasets, thereby markedly improving translation quality.

cs.CV↗

arXiv2024

Med-PerSAM: One-Shot Visual Prompt Tuning for Personalized Segment Anything Model in Medical Domain

Leveraging pre-trained models with tailored prompts for in-context learning has proven highly effective in NLP tasks. Building on this success, recent studies have applied a similar approach to the Segment Anything Model (SAM) within a ``one-shot" framework, where only a single reference image and its label are employed. However, these methods face limitations in the medical domain, primarily due to SAM's essential requirement for visual prompts and the over-reliance on pixel similarity for generating them. This dependency may lead to (1) inaccurate prompt generation and (2) clustering of point prompts, resulting in suboptimal outcomes. To address these challenges, we introduce \textbf{Med-PerSAM}, a novel and straightforward one-shot framework designed for the medical domain. Med-PerSAM uses only visual prompt engineering and eliminates the need for additional training of the pretrained SAM or human intervention, owing to our novel automated prompt generation process. By integrating our lightweight warping-based prompt tuning model with SAM, we enable the extraction and iterative refinement of visual prompts, enhancing the performance of the pre-trained SAM. This advancement is particularly meaningful in the medical domain, where creating visual prompts poses notable challenges for individuals lacking medical expertise. Our model outperforms various foundational models and previous SAM-based approaches across diverse 2D medical imaging datasets.

cs.CV↗

arXiv2024

Leveraging the Power of MLLMs for Gloss-Free Sign Language Translation

Sign language translation (SLT) is a challenging task that involves translating sign language images into spoken language. For SLT models to perform this task successfully, they must bridge the modality gap and identify subtle variations in sign language components to understand their meanings accurately. To address these challenges, we propose a novel gloss-free SLT framework called Multimodal Sign Language Translation (MMSLT), which leverages the representational capabilities of off-the-shelf multimodal large language models (MLLMs). Specifically, we use MLLMs to generate detailed textual descriptions of sign language components. Then, through our proposed multimodal-language pre-training module, we integrate these description features with sign video features to align them within the spoken sentence space. Our approach achieves state-of-the-art performance on benchmark datasets PHOENIX14T and CSL-Daily, highlighting the potential of MLLMs to be utilized effectively in SLT. Code is available at https://github.com/hwjeon98/MMSLT.

cs.CV↗

arXiv2022

Effective and Efficient Core Computation in Signed Networks

With the proliferation of mobile technology and IT development, people can use social network services at any place and anytime. Among many social network mining problems, identifying cohesive subgraphs attract many attentions from different fields due to its numerous applications. Among many cohesive subgraph models, k-core is the most widely used model due to its simple and intuitive structure. In this paper, we formulate (p,n)-core in signed networks by extending k-core. (p,n)-core simultaneously guarantees sufficient internal positive edges and deficient internal negative edges. We formally prove that finding an exact (p,n)-core is NP-hard. Hence, we propose three efficient and effective algorithms to find a solution. Using real-world and synthetic networks, we demonstrate the superiority of our proposed algorithms.

cs.SI↗

arXiv2022

LUEM : Local User Engagement Maximization in Networks

Understanding a social network is a fundamental problem in social network analysis because of its numerous applications. Recently, user engagement in networks has received extensive attention from many research groups. However, most user engagement models focus on global user engagement to maximize (or minimize) the number of engaged users. In this study, we formulate the so-called Local User Engagement Maximization (LUEM) problem. We prove that the LUEM problem is NP-hard. To obtain high-quality results, we propose an approximation algorithm that incorporates a traditional hill-climbing method. To improve efficiency, we propose an efficient pruning strategy while maintaining effectiveness. In addition, by observing the relationship between the degree and user engagement, we propose an efficient heuristic algorithm that preserves effectiveness. Finally, we conducted extensive experiments on ten real-world networks to demonstrate the superiority of the proposed algorithms. We observed that the proposed algorithm achieved up to 605% more engaged users compared to the best baseline algorithms.

cs.SI↗

arXiv2022

OCSM : Finding Overlapping Cohesive Subgraphs with Minimum Degree

Cohesive subgraph discovery in a network is one of the fundamental problems and investigated for several decades. In this paper, we propose the Overlapping Cohesive Subgraphs with Minimum degree (OCSM) problem which combines three key concepts for OCSM : (i) edge-based overlapping, (ii) the minimum degree constraint, and (iii) the graph density. To the best of our knowledge, this is the first work to identify overlapping cohesive subgraphs with minimum degree by incorporating the graph density. Since the OCSM problem is NP-hard, we propose two algorithms: advanced peeling algorithm and seed-based expansion algorithm. Finally, we show the experimental study with real-world networks to demonstrate the effectiveness and efficiency of our proposed algorithms.

cs.SI↗

arXiv2020

DPM: A Novel Training Method for Physics-Informed Neural Networks in Extrapolation

We present a method for learning dynamics of complex physical processes described by time-dependent nonlinear partial differential equations (PDEs). Our particular interest lies in extrapolating solutions in time beyond the range of temporal domain used in training. Our choice for a baseline method is physics-informed neural network (PINN) [Raissi et al., J. Comput. Phys., 378:686--707, 2019] because the method parameterizes not only the solutions but also the equations that describe the dynamics of physical processes. We demonstrate that PINN performs poorly on extrapolation tasks in many benchmark problems. To address this, we propose a novel method for better training PINN and demonstrate that our newly enhanced PINNs can accurately extrapolate solutions in time. Our method shows up to 72% smaller errors than existing methods in terms of the standard L2-norm metric.

cs.LG↗

arXiv2014

Magnetic structure of the conductive triangular-lattice antiferromagnet PdCrO2

We performed neutron single crystal and synchrotron X-ray powder diffraction experiments in order to investigate the magnetic and crystal structures of the conductive layered triangular-lattice antiferromagnet PdCrO2 with a putative spin chirality, which contributes to an unconventional anomalous Hall effect. We revealed that the ground-state magnetic structure is a commensurate and nearly-coplanar 120-degrees spin structure. The 120-degrees plane in different Cr layers seem to tilt with one another, leading to a small noncoplanarity. Such a small but finite non-coplanar stacking of the 120-degrees planes gives rise to a finite scalar spin chirality, which may be responsible for the unconventional nature of the Hall effect of PdCrO2.

cond-mat.str-el↗

arXiv2014

Switching of Conducting Planes by Partial Dimer Formation in IrTe$_2$

Single-crystal X-ray diffraction was employed to study the structural-electronic phase transition of IrTe$_2$ at approximately 270 K. The low-temperature structure was found to be a triclinic (space group $P\bar{1}$) characterized by the partial formation of Ir$_2$ dimers in the triangular lattice of IrTe$_2$, resulting in a structural modulation with a wave vector of ${\bf q} = (1/5, 0, -1/5)$. First-principles band calculations demonstrate that tilted two-dimensional Fermi surfaces emerge in the triclinic phase, suggesting that switching of the conducting planes occurs from the basal plane of the trigonal IrTe$_2$ to the tilted plane normal to ${\bf q}$ of the triclinic IrTe$_2$.

cond-mat.str-el↗

arXiv2013

Superconductivity Induced by Breaking Te2 Dimers of AuTe2

Mineral calaverite AuTe2 is a layered compound with an incommensurately modulated structure. The modulation is characterized by the formation of molecular-like Te2 dimers. We have found that the breaking of Te2 dimers that occurs in Au1-xPtxTe2 results in the emergence of superconductivity at Tc = 4.0 K.

cond-mat.supr-con↗

arXiv2012

Ferroelectricity Driven by Twisting of Silicate Tetrahedral Chains

Conventional perovskite-type ferroelectrics are based on octahedral units of oxygen, and often comprise toxic Pb to achieve robust ferroelectricity. Here, we report the ferroelectricity in a silicate-based compound, Bi2SiO5 (BSO), induced by a structural instability of the corresponding silicate tetrahedral chains. A low-energy phonon mode condenses at ~ 673 K to induce the proper ferroelectric phase transition. Polarization switching was observed in a BSO single crystal with a coercive field of 30 kV/cm and a spontaneous polarization of 0.3 microC/cm2 along a direction normal to the cleavage plane. The in-plane polarization was estimated by first principles calculations to be 23 microC/cm2. The present findings provide a new guideline for designing ferroelectric materials based on SiO4 tetrahedral units, which is ubiquitously found in natural minerals.

cond-mat.mtrl-sci↗