arXiv subjects

Dongjin Kim

Publications and source records attributed to Dongjin Kim.

29 records · Page 2Linked to original sources

arXiv2025

Machine learning interatomic potential can infer electrical response

Modeling the response of material and chemical systems to electric fields remains a longstanding challenge. Machine learning interatomic potentials (MLIPs) offer an efficient and scalable alternative to quantum mechanical methods but do not by themselves incorporate electrical response. Here, we show that polarization and Born effective charge (BEC) tensors can be directly extracted from long-range MLIPs within the Latent Ewald Summation (LES) framework, solely by learning from energy and force data. Using this approach, we predict the infrared spectra of bulk water under zero or finite external electric fields, ionic conductivities of high-pressure superionic ice, and the phase transition and hysteresis in ferroelectric PbTiO$_3$ perovskite. This work thus extends the capability of MLIPs to predict electrical response--without training on charges or polarization or BECs--and enables accurate modeling of electric-field-driven processes in diverse systems at scale.

cond-mat.mtrl-sci↗

arXiv2025

Watch Video, Catch Keyword: Context-aware Keyword Attention for Moment Retrieval and Highlight Detection

The goal of video moment retrieval and highlight detection is to identify specific segments and highlights based on a given text query. With the rapid growth of video content and the overlap between these tasks, recent works have addressed both simultaneously. However, they still struggle to fully capture the overall video context, making it challenging to determine which words are most relevant. In this paper, we present a novel Video Context-aware Keyword Attention module that overcomes this limitation by capturing keyword variation within the context of the entire video. To achieve this, we introduce a video context clustering module that provides concise representations of the overall video context, thereby enhancing the understanding of keyword dynamics. Furthermore, we propose a keyword weight detection module with keyword-aware contrastive learning that incorporates keyword information to enhance fine-grained alignment between visual and textual features. Extensive experiments on the QVHighlights, TVSum, and Charades-STA benchmarks demonstrate that our proposed method significantly improves performance in moment retrieval and highlight detection tasks compared to existing approaches. Our code is available at: https://github.com/VisualAIKHU/Keyword-DETR

cs.CV↗

arXiv2024

Learning charges and long-range interactions from energies and forces

Accurate modeling of long-range forces is critical in atomistic simulations, as they play a central role in determining the properties of materials and chemical systems. However, standard machine learning interatomic potentials (MLIPs) often rely on short-range approximations, limiting their applicability to systems with significant electrostatics and dispersion forces. We recently introduced the Latent Ewald Summation (LES) method, which captures long-range electrostatics without explicitly learning atomic charges or charge equilibration. Extending LES, we incorporate the ability to learn physical partial charges, encode charge states, and the option to impose charge neutrality constraints. We benchmark LES on diverse and challenging systems, including charged molecules, ionic liquid, electrolyte solution, polar dipeptides, surface adsorption, electrolyte/solid interfaces, and solid-solid interfaces. Our results show that LES can effectively infer physical partial charges, dipole and quadrupole moments, as well as achieve better accuracy compared to methods that explicitly learn charges. LES thus provides an efficient, interpretable, and generalizable MLIP framework for simulating complex systems with intricate charge transfer and long-range

physics.comp-ph↗

arXiv2024

Harnessing Meta-Learning for Improving Full-Frame Video Stabilization

Video stabilization is a longstanding computer vision problem, particularly pixel-level synthesis solutions for video stabilization which synthesize full frames add to the complexity of this task. These techniques aim to stabilize videos by synthesizing full frames while enhancing the stability of the considered video. This intensifies the complexity of the task due to the distinct mix of unique motion profiles and visual content present in each video sequence, making robust generalization with fixed parameters difficult. In our study, we introduce a novel approach to enhance the performance of pixel-level synthesis solutions for video stabilization by adapting these models to individual input video sequences. The proposed adaptation exploits low-level visual cues accessible during test-time to improve both the stability and quality of resulting videos. We highlight the efficacy of our methodology of "test-time adaptation" through simple fine-tuning of one of these models, followed by significant stability gain via the integration of meta-learning techniques. Notably, significant improvement is achieved with only a single adaptation step. The versatility of the proposed algorithm is demonstrated by consistently improving the performance of various pixel-level synthesis models for video stabilization in real-world scenarios.

cs.CV↗

arXiv2024

Learning to Visually Localize Sound Sources from Mixtures without Prior Source Knowledge

The goal of the multi-sound source localization task is to localize sound sources from the mixture individually. While recent multi-sound source localization methods have shown improved performance, they face challenges due to their reliance on prior information about the number of objects to be separated. In this paper, to overcome this limitation, we present a novel multi-sound source localization method that can perform localization without prior knowledge of the number of sound sources. To achieve this goal, we propose an iterative object identification (IOI) module, which can recognize sound-making objects in an iterative manner. After finding the regions of sound-making objects, we devise object similarity-aware clustering (OSC) loss to guide the IOI module to effectively combine regions of the same object but also distinguish between different objects and backgrounds. It enables our method to perform accurate localization of sound-making objects without any prior knowledge. Extensive experimental results on the MUSIC and VGGSound benchmarks show the significant performance improvements of the proposed method over the existing methods for both single and multi-source. Our code is available at: https://github.com/VisualAIKHU/NoPrior_MultiSSL

cs.CV↗

arXiv2024

REPrune: Channel Pruning via Kernel Representative Selection

Channel pruning is widely accepted to accelerate modern convolutional neural networks (CNNs). The resulting pruned model benefits from its immediate deployment on general-purpose software and hardware resources. However, its large pruning granularity, specifically at the unit of a convolution filter, often leads to undesirable accuracy drops due to the inflexibility of deciding how and where to introduce sparsity to the CNNs. In this paper, we propose REPrune, a novel channel pruning technique that emulates kernel pruning, fully exploiting the finer but structured granularity. REPrune identifies similar kernels within each channel using agglomerative clustering. Then, it selects filters that maximize the incorporation of kernel representatives while optimizing the maximum cluster coverage problem. By integrating with a simultaneous training-pruning paradigm, REPrune promotes efficient, progressive pruning throughout training CNNs, avoiding the conventional train-prune-finetune sequence. Experimental results highlight that REPrune performs better in computer vision tasks than existing methods, effectively achieving a balance between acceleration ratio and performance retention.

cs.CV↗

arXiv2023

Task Agnostic Restoration of Natural Video Dynamics

In many video restoration/translation tasks, image processing operations are naïvely extended to the video domain by processing each frame independently, disregarding the temporal connection of the video frames. This disregard for the temporal connection often leads to severe temporal inconsistencies. State-Of-The-Art (SOTA) techniques that address these inconsistencies rely on the availability of unprocessed videos to implicitly siphon and utilize consistent video dynamics to restore the temporal consistency of frame-wise processed videos which often jeopardizes the translation effect. We propose a general framework for this task that learns to infer and utilize consistent motion dynamics from inconsistent videos to mitigate the temporal flicker while preserving the perceptual quality for both the temporally neighboring and relatively distant frames without requiring the raw videos at test time. The proposed framework produces SOTA results on two benchmark datasets, DAVIS and videvo.net, processed by numerous image processing applications. The code and the trained models are available at \url{https://github.com/MKashifAli/TARONVD}.

cs.CV↗

arXiv2023

Audio-Visual Spatial Integration and Recursive Attention for Robust Sound Source Localization

The objective of the sound source localization task is to enable machines to detect the location of sound-making objects within a visual scene. While the audio modality provides spatial cues to locate the sound source, existing approaches only use audio as an auxiliary role to compare spatial regions of the visual modality. Humans, on the other hand, utilize both audio and visual modalities as spatial cues to locate sound sources. In this paper, we propose an audio-visual spatial integration network that integrates spatial cues from both modalities to mimic human behavior when detecting sound-making objects. Additionally, we introduce a recursive attention network to mimic human behavior of iterative focusing on objects, resulting in more accurate attention regions. To effectively encode spatial information from both modalities, we propose audio-visual pair matching loss and spatial region alignment loss. By utilizing the spatial cues of audio-visual modalities and recursively focusing objects, our method can perform more robust sound source localization. Comprehensive experimental results on the Flickr SoundNet and VGG-Sound Source datasets demonstrate the superiority of our proposed method over existing approaches. Our code is available at: https://github.com/VisualAIKHU/SIRA-SSL

cs.MM↗

arXiv2021

Ultrafast carrier-lattice interactions and interlayer modulations of Bi2Se3 by X-ray free electron laser diffraction

As a 3D topological insulator, bismuth selenide (Bi2Se3) has potential applications for electrically and optically controllable magnetic and optoelectronic devices. How the carriers interact with lattice is important to understand the coupling with its topological phase. It is essential to measure with a time scale smaller than picoseconds for initial interaction. Here we use an X-ray free-electron laser to perform time-resolved diffraction to study ultrafast carrier-induced lattice contractions and interlayer modulations in Bi2Se3 thin films. The lattice contraction depends on the carrier concentration and is followed by an interlayer expansion accompanied by oscillations. Using density functional theory (DFT) and the Lifshitz model, the initial contraction can be explained by van der Waals force modulation of the confined free carrier layers. Band inversion, related to a topological phase transition, is modulated by the expansion of the interlayer distance. These results provide insight into instantaneous topological phases on ultrafast timescales.

cond-mat.mtrl-sci↗

arXiv2020

Flux-Pinning Behaviors and Mechanism According to Dopant Level in (Fe, Ti) Paticle-Doped MgB$_2$ Superconductor

We have studied flux-pinning effects of MgB$_2$ superconductor by doping (Fe, Ti) particles of which radius is 163 nm on average. 5 wt.\% (Fe, Ti) doped MgB$_2$ among the specimens showed the best field dependence of magnetization and 25 wt.\% one did the worst at 5 K . The difference of field dependence of magnetization of the two increased as temperature increased. Here we show experimental results of (Fe, Ti) particle-doped MgB$_2$ according to dopant level and the causes of the behaviors. Flux-pinning effect of volume defects-doped superconductor was modeled in ideal state. During the study, we had to divide M-H curve of volume defect-dominating superconductor as three discreet regions for analyzing flux pinning effects, which are diamagnetic increase region, $Δ$H=$Δ$B region, and diamagnetic decrease region. As a result, flux-pinning effects of volume defects decreased as dopant level increased over the optimal dopant level, which was caused by decrease of flux-pinning limit of a volume defect. And similar behaviors are obtained as dopant level decreased below the optimal dopant level, which was caused by the decreased number of volume defects. Comparing the theory with experimental results, deviations increased as dopant level increased over the optimal dopant level, whereas the two was well matched on less dopant level than the optimal dopant level. The behavior is considered to be caused by segregation of volume defects. On the other hand, the property of over-doped specimens dramatically decrease as temperature increases, which is caused by double decreases of flux-pinning limit of a volume defect and segregation effect.

cond-mat.supr-con↗

arXiv2018

Astrometrically registered maps of H2O and SiO masers toward VX Sagittarii

The supergiant VX Sagittarii is a strong emitter of both H2O and SiO masers. However, previous VLBI observations have been performed separately, which makes it difficult to spatially trace the outward transfer of the material consecutively. Here we present the astrometrically registered, simultaneous maps of 22.2 GHz H2O and 43.1/42.8/86.2/129.3 GHz SiO masers toward VX Sagittarii. The H2O masers detected above the dust-forming layers have an asymmetric distribution. The multi-transition SiO masers are nearly circular ring, suggesting spherically symmetric wind within a few stellar radii. These results provide the clear evidence that the asymmetry in the outflow is enhanced after the smaller molecular gas clump transform into the inhomogeneous dust layers. The 129.3 GHz maser arises from the outermost region compared to that of 43.1/42.8/86.2 GHz SiO masers. The ring size of the 129.3 GHz maser is maximized around the optical maximum, suggesting that radiative pumping is dominant.

astro-ph.SR↗