arXiv subjects

Jaehong Lee

Publications and source records attributed to Jaehong Lee.

12 recordsLinked to original sources

arXiv2026

CV-HoloSR: Hologram to hologram super-resolution through volume-upsampling three-dimensional scenes

Existing hologram super-resolution (HSR) methods primarily focus on angle-of-view expansion. Adapting them for volumetric spatial up-sampling introduces severe quadratic depth distortion, degrading 3D focal accuracy. We propose CV-HoloSR, a complex-valued HSR framework specifically designed to preserve physically consistent linear depth scaling during volume up-sampling. Built upon a Complex-Valued Residual Dense Network (CV-RDN) and optimized with a novel depth-aware perceptual reconstruction loss, our model effectively suppresses over-smoothing to recover sharp, high-frequency interference patterns. To support this, we introduce a comprehensive large-depth-range dataset with resolutions up to 4K. Furthermore, to overcome the inherent depth bias of pre-trained encoders when scaling to massive target volumes, we integrate a parameter-efficient fine-tuning strategy utilizing complex-valued Low-Rank Adaptation (LoRA). Extensive numerical and physical optical experiments demonstrate our method's superiority. CV-HoloSR achieves a 32% improvement in perceptual realism (LPIPS of 0.2001) over state-of-the-art baselines. Additionally, our tailored LoRA strategy requires merely 200 samples, reducing training time by over 75% (from 22.5 to 5.2 hours) while successfully adapting the pre-trained backbone to unseen depth ranges and novel display configurations.

cs.GR↗

arXiv2025

MCAQ-YOLO: Morphological Complexity-Aware Quantization for Efficient Object Detection with Curriculum Learning

Most neural network quantization methods apply uniform bit precision across spatial regions, disregarding the heterogeneous complexity inherent in visual data. This paper introduces MCAQ-YOLO, a practical framework for tile-wise spatial mixed-precision quantization in real-time object detectors. Morphological complexity--quantified through five complementary metrics (fractal dimension, texture entropy, gradient variance, edge density, and contour complexity)--is proposed as a signal-centric predictor of spatial quantization sensitivity. A calibration-time analysis design enables spatial bit allocation with only 0.3ms inference overhead, achieving 151 FPS throughput. Additionally, a curriculum-based training scheme that progressively increases quantization difficulty is introduced to stabilize optimization and accelerate convergence. On a construction safety equipment dataset exhibiting high morphological variability, MCAQ-YOLO achieves 85.6% mAP@0.5 with an average bit-width of 4.2 bits and a 7.6x compression ratio, outperforming uniform 4-bit quantization by 3.5 percentage points. Cross-dataset evaluation on COCO 2017 (+2.9%) and Pascal VOC 2012 (+2.3%) demonstrates consistent improvements, with performance gains correlating with within-image complexity variation.

cs.CV↗

arXiv2025

A Large-Depth-Range Layer-Based Hologram Dataset for Machine Learning-Based 3D Computer-Generated Holography

Machine learning-based computer-generated holography (ML-CGH) has advanced rapidly in recent years, yet progress is constrained by the limited availability of high-quality, large-scale hologram datasets. To address this, we present KOREATECH-CGH, a publicly available dataset comprising 6,000 pairs of RGB-D images and complex holograms across resolutions ranging from 256*256 to 2048*2048, with depth ranges extending to the theoretical limits of the angular spectrum method for wide 3D scene coverage. To improve hologram quality at large depth ranges, we introduce amplitude projection, a post-processing technique that replaces amplitude components of hologram wavefields at each depth layer while preserving phase. This approach enhances reconstruction fidelity, achieving 27.01 dB PSNR and 0.87 SSIM, surpassing a recent optimized silhouette-masking layer-based method by 2.03 dB and 0.04 SSIM, respectively. We further validate the utility of KOREATECH-CGH through experiments on hologram generation and super-resolution using state-of-the-art ML models, confirming its applicability for training and evaluating next-generation ML-CGH systems.

cs.CV↗

arXiv2025

KoSimpleQA: A Korean Factuality Benchmark with an Analysis of Reasoning LLMs

We present $\textbf{Korean SimpleQA (KoSimpleQA)}$, a benchmark for evaluating factuality in large language models (LLMs) with a focus on Korean cultural knowledge. KoSimpleQA is designed to be challenging yet easy to grade, consisting of 1,000 short, fact-seeking questions with unambiguous answers. We conduct a comprehensive evaluation across a diverse set of open-source LLMs of varying sizes that support Korean, and find that even the strongest model generates correct answer only 33.7% of the time, underscoring the challenging nature of KoSimpleQA. Notably, performance rankings on KoSimpleQA differ substantially from those on the English SimpleQA, highlighting the unique value of our dataset. Furthermore, our analysis of reasoning LLMs shows that engaging reasoning capabilities in the factual QA task can both help models better elicit their latent knowledge and improve their ability to abstain when uncertain. KoSimpleQA can be found at https://anonymous.4open.science/r/KoSimpleQA-62EB.

cs.CL↗

arXiv2025

ReGUIDE: Data Efficient GUI Grounding via Spatial Reasoning and Search

Recent advances in Multimodal Large Language Models (MLLMs) have enabled autonomous agents to interact with computers via Graphical User Interfaces (GUIs), where accurately localizing the coordinates of interface elements (e.g., buttons) is often required for fine-grained actions. However, this remains significantly challenging, leading prior works to rely on large-scale web datasets to improve the grounding accuracy. In this work, we propose Reasoning Graphical User Interface Grounding for Data Efficiency (ReGUIDE), a novel and effective framework for web grounding that enables MLLMs to learn data efficiently through self-generated reasoning and spatial-aware criticism. More specifically, ReGUIDE learns to (i) self-generate a language reasoning process for the localization via online reinforcement learning, and (ii) criticize the prediction using spatial priors that enforce equivariance under input transformations. At inference time, ReGUIDE further boosts performance through a test-time scaling strategy, which combines spatial search with coordinate aggregation. Our experiments demonstrate that ReGUIDE significantly advances web grounding performance across multiple benchmarks, outperforming baselines with substantially fewer training data points (e.g., only 0.2% samples compared to the best open-sourced baselines).

cs.LG↗

arXiv2025

RT-HDIST: Ray-Tracing Core-based Hausdorff Distance Computation

The Hausdorff distance is a fundamental metric with widespread applications across various fields. However, its computation remains computationally expensive, especially for large-scale datasets. In this work, we present RT-HDIST, the first Hausdorff distance algorithm accelerated by ray-tracing cores (RT-cores). By reformulating the Hausdorff distance problem as a series of nearest-neighbor searches and introducing a novel quantized index space, RT-HDIST achieves significant reductions in computational overhead while maintaining exact results. Extensive benchmarks demonstrate up to a two-order-of-magnitude speedup over prior state-of-the-art methods, underscoring RT-HDIST's potential for real-time and large-scale applications.

cs.GR↗

arXiv2024

Paralinguistics-Aware Speech-Empowered Large Language Models for Natural Conversation

Recent work shows promising results in expanding the capabilities of large language models (LLM) to directly understand and synthesize speech. However, an LLM-based strategy for modeling spoken dialogs remains elusive, calling for further investigation. This paper introduces an extensive speech-text LLM framework, the Unified Spoken Dialog Model (USDM), designed to generate coherent spoken responses with naturally occurring prosodic features relevant to the given input speech without relying on explicit automatic speech recognition (ASR) or text-to-speech (TTS) systems. We have verified the inclusion of prosody in speech tokens that predominantly contain semantic information and have used this foundation to construct a prosody-infused speech-text model. Additionally, we propose a generalized speech-text pretraining scheme that enhances the capture of cross-modal semantics. To construct USDM, we fine-tune our speech-text model on spoken dialog data using a multi-step spoken dialog template that stimulates the chain-of-reasoning capabilities exhibited by the underlying LLM. Automatic and human evaluations on the DailyTalk dataset demonstrate that our approach effectively generates natural-sounding spoken responses, surpassing previous and cascaded baselines. Our code and checkpoints are available at https://github.com/naver-ai/usdm.

cs.CL↗

arXiv2024

HyperCLOVA X Technical Report

We introduce HyperCLOVA X, a family of large language models (LLMs) tailored to the Korean language and culture, along with competitive capabilities in English, math, and coding. HyperCLOVA X was trained on a balanced mix of Korean, English, and code data, followed by instruction-tuning with high-quality human-annotated datasets while abiding by strict safety guidelines reflecting our commitment to responsible AI. The model is evaluated across various benchmarks, including comprehensive reasoning, knowledge, commonsense, factuality, coding, math, chatting, instruction-following, and harmlessness, in both Korean and English. HyperCLOVA X exhibits strong reasoning capabilities in Korean backed by a deep understanding of the language and cultural nuances. Further analysis of the inherent bilingual nature and its extension to multilingualism highlights the model's cross-lingual proficiency and strong generalization ability to untargeted languages, including machine translation between several language pairs and cross-lingual inference tasks. We believe that HyperCLOVA X can provide helpful guidance for regions or countries in developing their sovereign LLMs.

cs.CL↗

arXiv2023

Analysis and active control of geometrically nonlinear responses of smart FG porous plates with graphene nanoplatelets reinforcement based on Bézier extraction of NURBS

In this paper, we propose an effective computational approach to analyze and active control of geometrically nonlinear responses of functionally graded (FG) porous plates with graphene nanoplatelets (GPLs) reinforcement integrated with piezoelectric layers. The key concept behind this work is to utilize isogeometric analysis (IGA) based on Bézier extraction technique and $C^0$-type higher-order shear deformation theory ($C^0$-HSDT). By applying Bézier extraction, the original Non-Uniform Rational B-Spline (NURBS) control meshes can be transformed into Bézier elements which allow us to inherit the standard numerical procedure like the standard finite element method (FEM). In this scenario, the approximation of mechanical displacement field is calculated via $C^0$-HSDT whilst the electric potential field is considered as a linear function across the thickness of each piezoelectric sublayer. The FG plate includes internal pores and GPLs dispersed into metal matrix either uniformly or non-uniformly along plate's thickness. To control responses of structures, the top and bottom surfaces of FG plate are firmly bonded with piezoelectric layers which are considered as sensor and actuator layers. The geometrically nonlinear equations are solved by Newton-Raphson iterative procedure and Newmark's integration. The influence of porosity coefficient, weight fraction of GPLs as well as external electrical voltage on geometrically nonlinear behaviors of plate structures with various distributions of porosity and GPLs are thoroughly investigated. A constant displacement and velocity feedback control approaches are then adopted to actively control geometrically nonlinear static and dynamic responses, where structural damping effect is taken into account, based on a closed-loop control with sensor and actuator layers.

physics.app-ph↗

arXiv2023

Modelling of FG-TPMS plates

Functionally graded porous plates have been validated as remarkable lightweight structures with excellent mechanical characteristics and numerous applications. With inspiration from the high strength-to-volume ratio of triply periodic minimal surface (TPMS) structures, a new model of porous plates, which is called a functionally graded TPMS (FG-TPMS) plate, is investigated in this paper. Three TPMS architectures including Primitive (P), Gyroid (G), and wrapped package-graph (IWP) with different graded functions are presented. To predict the mechanical responses, a new fitting technique based on a two-phase piece-wise function is employed to evaluate the effective moduli of TPMS structures, including elastic modulus, shear modulus, and bulk modulus. In addition, this function corresponds to the cellular structure formulation in the context of relative density. The separated phases of the function are divided by the different deformation behaviors. Furthermore, another crucial mechanical property of porous structure, i.e, Poisson's ratio, is also achieved by a similar fitting technique. To verify the mechanical characteristics of the FG-TPMS plate, the generalized displacement field is modeled by a seventh-order shear deformation theory (SeSDT) and isogeometric analysis (IGA). Numerical examples regarding static, buckling, and free vibration analyses of FG-TPMS plates are illustrated to confirm the reliability and accuracy of the proposed approach. Consequently, these FG-TPMS structures can provide much higher stiffness than the same-weight isotropic plate. The greater stiffness-to-weight ratio of these porous plates compared to the full-weight isotropic ones should be considered the most remarkable feature. Thus, these complex porous structures have numerous practical applications because of these high ratios and their fabrication ability through additive manufacturing (AM) technology.

cond-mat.mtrl-sci↗

arXiv2015

Isogeometric approach for nonlinear bending and post-buckling analysis of functionally graded plates under thermal environment

In this paper, equilibrium and stability equations of functionally graded material (FGM) plate under thermal environment are formulated based on isogeometric analysis (IGA) in combination with higher-order shear deformation theory (HSDT). The FGM plate is made by a mixture of two distinct components, for which material properties not only vary continuously through thickness according to a power-law distribution but also are assumed to be a function of temperature. Temperature field is assumed to be constant in any plane and uniform, linear and nonlinear through plate thickness, respectively. The governing equation is in nonlinear form based on von Karman assumption and thermal effect. A NURBS-based isogeometric finite element formulation is utilized to naturally fulfil the rigorous C1-continuity required by the present plate model. Influences of gradient indices, boundary conditions, temperature distributions, material properties, length-to-thickness ratios on the behaviour of FGM plate are discussed in details. Numerical results demonstrate excellent performance of the present approach.

cs.CE↗

arXiv2014

Geometrically nonlinear isogeometric analysis of laminated composite plates based on higher-order shear deformation theory

In this paper, we present an effectively numerical approach based on isogeometric analysis (IGA) and higher-order shear deformation theory (HSDT) for geometrically nonlinear analysis of laminated composite plates. The HSDT allows us to approximate displacement field that ensures by itself the realistic shear strain energy part without shear correction factors. IGA utilizing basis functions namely B-splines or non-uniform rational B-splines (NURBS) enables to satisfy easily the stringent continuity requirement of the HSDT model without any additional variables. The nonlinearity of the plates is formed in the total Lagrange approach based on the von-Karman strain assumptions. Numerous numerical validations for the isotropic, orthotropic, cross-ply and angle-ply laminated plates are provided to demonstrate the effectiveness of the proposed method.

cs.CE↗