arXiv subjects

Minjung Kim

Publications and source records attributed to Minjung Kim.

At least 19 recordsLinked to original sources

arXiv2026

Investigation of hadronic effects on resonance productions in small collision systems using the EPOS4 model

Recent experimental results in high-multiplicity proton-proton (pp) collisions have suggested the possible emergence of collective behavior and medium-like effects previously considered characteristic of heavy-ion collisions. Resonance production provides a sensitive probe of such effects, as resonance yields and transverse-momentum distributions can be modified by hadronic interactions occurring between chemical and kinetic freeze-out. In this study, these effects are investigated using the EPOS4 event generator, in which hadronic final-state interactions are modeled through the UrQMD transport approach. By comparing calculations performed with and without UrQMD, the impact of hadronic interactions on resonance production is evaluated. In addition, the UrQMD contributions are separated into regeneration and rescattering, enabling a detailed investigation of both resonance production enhancement and the loss of reconstructible resonance signals. The analysis is performed for various mesonic and baryonic resonances with different lifetimes in pp collisions at LHC energies and is extended to p-O, O-O, and Pb-Pb collisions to study the system-size dependence of hadronic-phase effects. The results show that resonance production is governed by the competition between regeneration and rescattering, whose relative importance depends strongly on the resonance species, transverse momentum, and collision system. While rescattering suppresses reconstructible short-lived resonance signals, regeneration can significantly enhance the yields of several resonance species, particularly baryonic resonances. These findings demonstrate that hadronic interactions can play an important role even in small collision systems and highlight the need to measure resonances with different lifetimes and quantum numbers to constrain the dynamics and lifetime of the hadronic phase across collision systems.

nucl-ex↗

arXiv2026

OTT-Vid: Optimal Transport Temporal Token Compression for Video Large Language Models

As Video Large Language Models (Video-LLMs) scale to longer and more complex videos, their inference cost grows rapidly due to the large volume of visual tokens accumulated across frames. Training-free token compression has emerged as a practical solution to this bottleneck. However, existing temporal compression methods rely primarily on cross-frame token similarity or segmentation heuristics, overlooking each token's semantic role within its frame and failing to adapt compression strength to the compressibility of each frame pair. In this work, we propose OTT-Vid, a transport-derived allocation framework for temporal token compression. Our approach consists of two stages: spatial pruning identifies representative content within each frame, and optimal transport (OT) is then solved between neighboring frames to estimate temporal compressibility. We formulate this OT with non-uniform token mass, which protects semantically important tokens from aggressive compression, and a locality-aware cost that captures both feature and spatial disparities. The resulting transport plan jointly balances token importance and matching cost, while its total cost defines the transport difficulty of each frame pair, which we use to allocate compression budgets dynamically. Experiments on six benchmarks spanning video question answering and temporal grounding show that OTT-Vid preserves 95.8% of VQA and 73.9% of VTG performance while retaining only 10% of tokens, consistently outperforming existing state-of-the-art training-free compression methods.

cs.CV↗

arXiv2026

Reading Speed, Image Quality Ratings, and Comfort Ratings in Augmented Reality

The rendering and display of text is a key use-case for augmented reality (AR). Here, we present the Read-AR, a dataset of reading in AR, for which we collected over 11,000 reading speeds and almost 6000 visual quality and comfort ratings across over 80 different experiment conditions on the same experiment set-up. The consistent, controlled set-up enables the dataset to function as a reference for benchmarking the quality of different AR headset architectures.

cs.HC↗

arXiv2026

Revisiting Weakly-Supervised Video Scene Graph Generation via Pair Affinity Learning

Weakly-supervised video scene graph generation (WS-VSGG) aims to parse video content into structured relational triplets without bounding box annotations and with only sparse temporal labeling, significantly reducing annotation costs. Without ground-truth bounding boxes, these methods rely on off-the-shelf detectors to generate object proposals, yet largely overlook a fundamental discrepancy from fullysupervised pipelines. Fully-supervised detectors implicitly filter out noninteractive objects, while off-the-shelf detectors indiscriminately detect all visible objects, overwhelming relation models with noisy pairs.We address this by introducing a learnable pair affinity that estimates the likelihood of interaction between subject-object pairs. Through Pair Affinity Learning and Scoring (PALS), pair affinity is incorporated into inferencetime ranking and further integrated into contextual reasoning through Pair Affinity Modulation (PAM), enabling the model to suppress noninteractive pairs and focus on relationally meaningful ones. To provide cleaner supervision for pair affinity learning, we further propose Relation- Aware Matching (RAM), which leverages vision-language grounding to resolve class-level ambiguity in pseudo-label generation. Extensive experiments on Action Genome demonstrate that our approach consistently yields substantial improvements across different baselines and backbones, achieving state-of-the-art WS-VSGG performance.

cs.CV↗

arXiv2026

MAPLE: Modality-Aware Post-training and Learning Ecosystem

Multimodal language models now integrate text, audio, and video for unified reasoning. Yet existing RL post-training pipelines treat all input signals as equally relevant, ignoring which modalities each task actually requires. This modality-blind training inflates policy-gradient variance, slows convergence, and degrades robustness to real-world distribution shifts where signals may be missing, added, or reweighted. We introduce MAPLE, a complete modality-aware post-training and learning ecosystem comprising: (1) MAPLE-bench, the first benchmark explicitly annotating minimal signal combinations required per task; (2) MAPO, a modality-aware policy optimization framework that stratifies batches by modality requirement to reduce gradient variance from heterogeneous group advantages; (3) Adaptive weighting and curriculum scheduling that balances and prioritizes harder signal combinations. Systematic analysis across loss aggregation, clipping, sampling, and curriculum design establishes MAPO's optimal training strategy. Adaptive weighting and curriculum focused learning further boost performance across signal combinations. MAPLE narrows uni/multi-modal accuracy gaps by 30.24%, converges 3.18x faster, and maintains stability across all modality combinations under realistic reduced signal access. MAPLE constitutes a complete recipe for deployment-ready multimodal RL post-training.

cs.AI↗

arXiv2025

TSGM: Regular and Irregular Time-series Generation using Score-based Generative Models

Score-based generative models (SGMs) have demonstrated unparalleled sampling quality and diversity in numerous fields, such as image generation, voice synthesis, and tabular data synthesis, etc. Inspired by those outstanding results, we apply SGMs to synthesize time-series by learning its conditional score function. To this end, we present a conditional score network for time-series synthesis, deriving a denoising score matching loss tailored for our purposes. In particular, our presented denoising score matching loss is the conditional denoising score matching loss for time-series synthesis. In addition, our framework is such flexible that both regular and irregular time-series can be synthesized with minimal changes to our model design. Finally, we obtain exceptional synthesis performance on various time-series datasets, achieving state-of-the-art sampling diversity and quality.

cs.LG↗

arXiv2025

Empower Words: DualGround for Structured Phrase and Sentence-Level Temporal Grounding

Video Temporal Grounding (VTG) aims to localize temporal segments in long, untrimmed videos that align with a given natural language query. This task typically comprises two subtasks: Moment Retrieval (MR) and Highlight Detection (HD). While recent advances have been progressed by powerful pretrained vision-language models such as CLIP and InternVideo2, existing approaches commonly treat all text tokens uniformly during crossmodal attention, disregarding their distinct semantic roles. To validate the limitations of this approach, we conduct controlled experiments demonstrating that VTG models overly rely on [EOS]-driven global semantics while failing to effectively utilize word-level signals, which limits their ability to achieve fine-grained temporal alignment. Motivated by this limitation, we propose DualGround, a dual-branch architecture that explicitly separates global and local semantics by routing the [EOS] token through a sentence-level path and clustering word tokens into phrase-level units for localized grounding. Our method introduces (1) tokenrole- aware cross modal interaction strategies that align video features with sentence-level and phrase-level semantics in a structurally disentangled manner, and (2) a joint modeling framework that not only improves global sentence-level alignment but also enhances finegrained temporal grounding by leveraging structured phrase-aware context. This design allows the model to capture both coarse and localized semantics, enabling more expressive and context-aware video grounding. DualGround achieves state-of-the-art performance on both Moment Retrieval and Highlight Detection tasks across QVHighlights and Charades- STA benchmarks, demonstrating the effectiveness of disentangled semantic modeling in video-language alignment.

cs.CV↗

arXiv2025

Further Characterisation of Digital Pixel Test Structures Implemented in a 65 nm CMOS Process

The next generation of MAPS for future tracking detectors will have to meet stringent requirements placed on them. One such detector is the ALICE ITS3 that aims to be very light at 0.07% X/X$_{0}$ per layer and have a low power consumption in the active area of 40 mW/cm$^{2}$ by implementing wafer-scale MAPS bent into cylindrical half layers. To address these challenging requirements, the ALICE ITS3 project, in conjunction with the CERN EP R&D on monolithic pixel sensors, proposed the Tower Partners Semiconductor Co. 65 nm CMOS process as the starting point for the sensor. After the initial results confirmed the detection efficiency and radiation hardness, the choice of the technology was solidified by demonstrating the feasibility of operating MAPS in low-power consumption regimes, < 50 mW/cm$^{2}$, while maintaining high-quality performance. This was shown through a detailed characterisation of the Digital Pixel Test Structure (DPTS) prototype exposed to X-rays and ionising beams, and the results are presented in this article. Additionally, the sensor was further investigated through studies of the fake-hit rate, the linearity of the front-end in the range 1.7-28 keV, the performance after ionising irradiation, and the detection efficiency of inclined tracks in the range 0-45$^\circ$.

physics.ins-det↗

arXiv2025

Photoproduction and detection of $\rho'\rightarrow\pi^+\pi^-\pi^+\pi^-$ decays in ultra-peripheral collisions and at an electron-ion collider

Vector meson photoproduction is an important probe of nuclear structure. Light vector mesons are most sensitive to low$-x$ structure, as long as they are not too light for perturbative QCD calculations. The $\rho'$ is of interest as an intermediate mass state (between the $\rho$ and $J/\psi$) that is easier to detect than the $\phi$. Using HERA data on proton targets, we make projections for lead/gold targets in UPCs at the Large Hadron Collider and RHIC, and for $ep$ and $eA$ collisions at a future Electron-Ion Collider (EIC). We compare the UPC projections with ALICE data, and constrain the branching ratio divided by the square of the photon-$\rho'$ coupling. The data prefer large couplings and small branching ratio, probably less than 25\%. The photon-meson coupling predicted by generalized vector meson dominance does not fit the data. The HERA $ep$ and ALICE UPC $e$Pb data exhibit very similar $4\pi$ mass spectra, indicating that, if the system is composed of two resonances, the products of their photon couplings with their four-pion branching ratios are similar. The predicted rates are high for both UPCs and the EIC. The $\rho'\rightarrow\pi^+\pi^-\pi^+\pi^-$ decay can be observed at the EIC with high efficiency. In $ep$ collisions at the highest energy, the forward B0 detector is needed to observe this channel down to the lowest achievable Bjorken$-x$ values.

hep-ph↗

arXiv2025

Possibility for Proactive Anomaly Detection

Time-series anomaly detection, which detects errors and failures in a workflow, is one of the most important topics in real-world applications. The purpose of time-series anomaly detection is to reduce potential damages or losses. However, existing anomaly detection models detect anomalies through the error between the model output and the ground truth (observed) value, which makes them impractical. In this work, we present a \textit{proactive} approach for time-series anomaly detection based on a time-series forecasting model specialized for anomaly detection and a data-driven anomaly detection model. Our proactive approach establishes an anomaly threshold from training data with a data-driven anomaly detection model, and anomalies are subsequently detected by identifying predicted values that exceed the anomaly threshold. In addition, we extensively evaluated the model using four anomaly detection benchmarks and analyzed both predictable and unpredictable anomalies. We attached the source code as supplementary material.

cs.LG↗

arXiv2024

Neural Fourier Modelling: A Highly Compact Approach to Time-Series Analysis

Neural time-series analysis has traditionally focused on modeling data in the time domain, often with some approaches incorporating equivalent Fourier domain representations as auxiliary spectral features. In this work, we shift the main focus to frequency representations, modeling time-series data fully and directly in the Fourier domain. We introduce Neural Fourier Modelling (NFM), a compact yet powerful solution for time-series analysis. NFM is grounded in two key properties of the Fourier transform (FT): (i) the ability to model finite-length time series as functions in the Fourier domain, treating them as continuous-time elements in function space, and (ii) the capacity for data manipulation (such as resampling and timespan extension) within the Fourier domain. We reinterpret Fourier-domain data manipulation as frequency extrapolation and interpolation, incorporating this as a core learning mechanism in NFM, applicable across various tasks. To support flexible frequency extension with spectral priors and effective modulation of frequency representations, we propose two learning modules: Learnable Frequency Tokens (LFT) and Implicit Neural Fourier Filters (INFF). These modules enable compact and expressive modeling in the Fourier domain. Extensive experiments demonstrate that NFM achieves state-of-the-art performance on a wide range of tasks (forecasting, anomaly detection, and classification), including challenging time-series scenarios with previously unseen sampling rates at test time. Moreover, NFM is highly compact, requiring fewer than 40K parameters in each task, with time-series lengths ranging from 100 to 16K.

cs.LG↗

arXiv2024

See It All: Contextualized Late Aggregation for 3D Dense Captioning

3D dense captioning is a task to localize objects in a 3D scene and generate descriptive sentences for each object. Recent approaches in 3D dense captioning have adopted transformer encoder-decoder frameworks from object detection to build an end-to-end pipeline without hand-crafted components. However, these approaches struggle with contradicting objectives where a single query attention has to simultaneously view both the tightly localized object regions and contextual environment. To overcome this challenge, we introduce SIA (See-It-All), a transformer pipeline that engages in 3D dense captioning with a novel paradigm called late aggregation. SIA simultaneously decodes two sets of queries-context query and instance query. The instance query focuses on localization and object attribute descriptions, while the context query versatilely captures the region-of-interest of relationships between multiple objects or with the global scene, then aggregated afterwards (i.e., late aggregation) via simple distance-based measures. To further enhance the quality of contextualized caption generation, we design a novel aggregator to generate a fully informed caption based on the surrounding context, the global environment, and object instances. Extensive experiments on two of the most widely-used 3D dense captioning datasets demonstrate that our proposed method achieves a significant improvement over prior methods.

cs.CV↗

arXiv2024

Bi-directional Contextual Attention for 3D Dense Captioning

3D dense captioning is a task involving the localization of objects and the generation of descriptions for each object in a 3D scene. Recent approaches have attempted to incorporate contextual information by modeling relationships with object pairs or aggregating the nearest neighbor features of an object. However, the contextual information constructed in these scenarios is limited in two aspects: first, objects have multiple positional relationships that exist across the entire global scene, not only near the object itself. Second, it faces with contradicting objectives--where localization and attribute descriptions are generated better with tight localization, while descriptions involving global positional relations are generated better with contextualized features of the global scene. To overcome this challenge, we introduce BiCA, a transformer encoder-decoder pipeline that engages in 3D dense captioning for each object with Bi-directional Contextual Attention. Leveraging parallelly decoded instance queries for objects and context queries for non-object contexts, BiCA generates object-aware contexts, where the contexts relevant to each object is summarized, and context-aware objects, where the objects relevant to the summarized object-aware contexts are aggregated. This extension relieves previous methods from the contradicting objectives, enhancing both localization performance and enabling the aggregation of contextual features throughout the global scene; thus improving caption generation performance simultaneously. Extensive experiments on two of the most widely-used 3D dense captioning datasets demonstrate that our proposed method achieves a significant improvement over prior methods.

cs.CV↗

arXiv2024

Marvelous Agglutinative Language Effect on Cross Lingual Transfer Learning

As for multilingual language models, it is important to select languages for training because of the curse of multilinguality. It is known that using languages with similar language structures is effective for cross lingual transfer learning. However, we demonstrate that using agglutinative languages such as Korean is more effective in cross lingual transfer learning. This is a great discovery that will change the training strategy of cross lingual transfer learning.

cs.CL↗

arXiv2024

K+K- photoproduction in ultra$-$peripheral Pb$--$Pb collisions

In ultra-peripheral collisions (UPCs) of relativistic heavy ions, photoproduction occurs when a photon emitted from one nucleus interacts with the other nucleus from the opposing beam, producing particles in the final state. Measurements of \KK photoproduction probe interactions and couplings between the $ϕ(1020)$ and charged kaons with photons and nuclear targets. We report exclusive \KK photoproduction cross section at midrapidity in \PbPb collisions at \snn = 5.02 TeV, which is measured for the first time in UPCs.

nucl-ex↗

arXiv2024

FIMP: Future Interaction Modeling for Multi-Agent Motion Prediction

Multi-agent motion prediction is a crucial concern in autonomous driving, yet it remains a challenge owing to the ambiguous intentions of dynamic agents and their intricate interactions. Existing studies have attempted to capture interactions between road entities by using the definite data in history timesteps, as future information is not available and involves high uncertainty. However, without sufficient guidance for capturing future states of interacting agents, they frequently produce unrealistic trajectory overlaps. In this work, we propose Future Interaction modeling for Motion Prediction (FIMP), which captures potential future interactions in an end-to-end manner. FIMP adopts a future decoder that implicitly extracts the potential future information in an intermediate feature-level, and identifies the interacting entity pairs through future affinity learning and top-k filtering strategy. Experiments show that our future interaction modeling improves the performance remarkably, leading to superior performance on the Argoverse motion forecasting benchmark.

cs.CV↗

arXiv2023

EP2P-Loc: End-to-End 3D Point to 2D Pixel Localization for Large-Scale Visual Localization

Visual localization is the task of estimating a 6-DoF camera pose of a query image within a provided 3D reference map. Thanks to recent advances in various 3D sensors, 3D point clouds are becoming a more accurate and affordable option for building the reference map, but research to match the points of 3D point clouds with pixels in 2D images for visual localization remains challenging. Existing approaches that jointly learn 2D-3D feature matching suffer from low inliers due to representational differences between the two modalities, and the methods that bypass this problem into classification have an issue of poor refinement. In this work, we propose EP2P-Loc, a novel large-scale visual localization method that mitigates such appearance discrepancy and enables end-to-end training for pose estimation. To increase the number of inliers, we propose a simple algorithm to remove invisible 3D points in the image, and find all 2D-3D correspondences without keypoint detection. To reduce memory usage and search complexity, we take a coarse-to-fine approach where we extract patch-level features from 2D images, then perform 2D patch classification on each 3D point, and obtain the exact corresponding 2D pixel coordinates through positional encoding. Finally, for the first time in this task, we employ a differentiable PnP for end-to-end training. In the experiments on newly curated large-scale indoor and outdoor benchmarks based on 2D-3D-S and KITTI, we show that our method achieves the state-of-the-art performance compared to existing visual localization and image-to-point cloud registration methods.

cs.CV↗

arXiv2023

MadSGM: Multivariate Anomaly Detection with Score-based Generative Models

The time-series anomaly detection is one of the most fundamental tasks for time-series. Unlike the time-series forecasting and classification, the time-series anomaly detection typically requires unsupervised (or self-supervised) training since collecting and labeling anomalous observations are difficult. In addition, most existing methods resort to limited forms of anomaly measurements and therefore, it is not clear whether they are optimal in all circumstances. To this end, we present a multivariate time-series anomaly detector based on score-based generative models, called MadSGM, which considers the broadest ever set of anomaly measurement factors: i) reconstruction-based, ii) density-based, and iii) gradient-based anomaly measurements. We also design a conditional score network and its denoising score matching loss for the time-series anomaly detection. Experiments on five real-world benchmark datasets illustrate that MadSGM achieves the most robust and accurate predictions.

cs.LG↗