arXiv subjects

Sanghyeon Kim

Publications and source records attributed to Sanghyeon Kim.

7 recordsLinked to original sources

arXiv2026

Distance-Aware Attention and Wall-Distance Expert Routing for Transformer-Based 3D Flow Prediction

Transformer surrogates for 3D flow prediction compress an industrial mesh into a small set of tokens from which every prediction point reads. Two operations follow: the retrieval step in which a point gathers information from the compressed representation, and the feed-forward layer that transforms what it retrieved. In current backbones both are blind to where the point sits in the flow. We condition both on wall-related physical signals. Distance-aware cross-attention (DA-CA) reshapes each volume query by its wall distance before retrieval, so that a point deep in the boundary layer draws different geometric information than one in the outer flow. Surface-volume mixture-of-experts (SVMoE) replaces the shared feed-forward layer with a small set of experts, routed by wall distance for volume points and by local geometry for surface points. Neither mechanism is tied to one architecture, so we apply both unchanged to AB-UPT and Transolver-3. On DrivAerML with 50 training cases, DA-CA reduces the volume pressure error by 10.1%, and DA-CA and SVMoE together reduce it by 12.5%; DA-CA improves the near-wall region at some cost in the far region, which SVMoE recovers, and the volume experts settle into near-wall, transition, and free-stream bands without routing supervision. Retrained on 300 cases, the conditioning improves every field quantity, reducing volume pressure and velocity errors by 33.1% and 18.6% on AB-UPT and by 21.4% and 21.3% on Transolver-3. Under Leave-One-Body-Out evaluation on DrivAerNet++, it reduces the volume pressure error on unseen body types by up to 14.2%.

cs.AI↗

arXiv2026

Universal zero-crosstalk photonic integration via slab-engineered mode hybridization

Photonic integrated circuits have emerged as a scalable platform for optical computing, communication, and quantum technologies, where high-fidelity optical processing is essential. However, as photonic systems scale in complexity, inter-channel crosstalk accumulates across cascaded components, fundamentally degrading signal fidelity, limiting system-level performance, and constraining integration density. Existing crosstalk-suppression strategies rely on specialized nanostructures or platform-specific designs, hindering their adoption in standard foundry processes and across diverse material systems. Here we establish a universal and foundry-compatible route to eliminating crosstalk based on slab-engineered mode hybridization in standard rib waveguides. By tailoring the slab thickness, mode hybridization induces anisotropic modal perturbations that enable complete cancellation of coupling between adjacent waveguides. We experimentally demonstrate zero-crosstalk across diverse material platforms, including silicon-on-insulator, silicon nitride, thin-film lithium niobate, and germanium-on-insulator, spanning wavelengths from the visible to the mid-infrared. Our approach provides a manufacturable route toward scalable, high-fidelity, and high-density photonic integration, overcoming the long-standing trade-off between signal fidelity and integration density in large-scale photonic systems.

physics.optics↗

arXiv2026

Generalized model of anisotropic thermo-optic response on thin-film lithium niobate platform

Thermo-optic (TO) control is crucial for thin-film lithium niobate (TFLN) photonic integrated circuits (PICs), offering a simple and practical method for low-frequency and DC tuning while remaining compatible with high-frequency electro-optic (EO) modulation. In x-cut TFLN, the TO response is inherently anisotropic, depending on both waveguide propagation angle and polarization due to the mode-specific overlap of the electric field with the ordinary and extraordinary refractive index axes of the crystal. Despite its significance, a systematic and quantitative analysis of this anisotropy has remained elusive. Here, we present the first generalized analytical model that describes the anisotropic TO response as a function of polarization and arbitrary waveguide orientation, and rigorously validate it through numerical simulations and experiments. This study provides foundational insight into anisotropic thermal tuning and enables new opportunities for engineering energy-efficient and scalable photonic design in next-generation TFLN PICs.

physics.optics↗

arXiv2024

Machine learning approach to brain tumor detection and classification

Brain tumor detection and classification are critical tasks in medical image analysis, particularly in early-stage diagnosis, where accurate and timely detection can significantly improve treatment outcomes. In this study, we apply various statistical and machine learning models to detect and classify brain tumors using brain MRI images. We explore a variety of statistical models including linear, logistic, and Bayesian regressions, and the machine learning models including decision tree, random forest, single-layer perceptron, multi-layer perceptron, convolutional neural network (CNN), recurrent neural network, and long short-term memory. Our findings show that CNN outperforms other models, achieving the best performance. Additionally, we confirm that the CNN model can also work for multi-class classification, distinguishing between four categories of brain MRI images such as normal, glioma, meningioma, and pituitary tumor images. This study demonstrates that machine learning approaches are suitable for brain tumor detection and classification, facilitating real-world medical applications in assisting radiologists with early and accurate diagnosis.

cs.CV↗

arXiv2023

Hydra: Multi-head Low-rank Adaptation for Parameter Efficient Fine-tuning

The recent surge in large-scale foundation models has spurred the development of efficient methods for adapting these models to various downstream tasks. Low-rank adaptation methods, such as LoRA, have gained significant attention due to their outstanding parameter efficiency and no additional inference latency. This paper investigates a more general form of adapter module based on the analysis that parallel and sequential adaptation branches learn novel and general features during fine-tuning, respectively. The proposed method, named Hydra, due to its multi-head computational branches, combines parallel and sequential branch to integrate capabilities, which is more expressive than existing single branch methods and enables the exploration of a broader range of optimal points in the fine-tuning process. In addition, the proposed adaptation method explicitly leverages the pre-trained weights by performing a linear combination of the pre-trained features. It allows the learned features to have better generalization performance across diverse downstream tasks. Furthermore, we perform a comprehensive analysis of the characteristics of each adaptation branch with empirical evidence. Through an extensive range of experiments, encompassing comparisons and ablation studies, we substantiate the efficiency and demonstrate the superior performance of Hydra. This comprehensive evaluation underscores the potential impact and effectiveness of Hydra in a variety of applications. Our code is available on \url{https://github.com/extremebird/Hydra}

cs.CV↗

arXiv2023

SMPConv: Self-moving Point Representations for Continuous Convolution

Continuous convolution has recently gained prominence due to its ability to handle irregularly sampled data and model long-term dependency. Also, the promising experimental results of using large convolutional kernels have catalyzed the development of continuous convolution since they can construct large kernels very efficiently. Leveraging neural networks, more specifically multilayer perceptrons (MLPs), is by far the most prevalent approach to implementing continuous convolution. However, there are a few drawbacks, such as high computational costs, complex hyperparameter tuning, and limited descriptive power of filters. This paper suggests an alternative approach to building a continuous convolution without neural networks, resulting in more computationally efficient and improved performance. We present self-moving point representations where weight parameters freely move, and interpolation schemes are used to implement continuous functions. When applied to construct convolutional kernels, the experimental results have shown improved performance with drop-in replacement in the existing frameworks. Due to its lightweight structure, we are first to demonstrate the effectiveness of continuous convolution in a large-scale setting, e.g., ImageNet, presenting the improvements over the prior arts. Our code is available on https://github.com/sangnekim/SMPConv

cs.CV↗

arXiv2021

Thermal conductivity of intercalation, conversion, and alloying lithium-ion battery electrode materials as function of their state of charge

Upon insertion and extraction of lithium, materials important for electrochemical energy storage can undergo changes in thermal conductivity ($Λ$) and elastic modulus ($\it M$). These changes are attributed to evolution of the intrinsic thermal carrier lifetime and interatomic bonding strength associated with structural transitions of electrode materials with varying degrees of reversibility. Using in situ time-domain thermoreflectance (TDTR) and picosecond acoustics, we systemically study $Λ$ and $\it M$ of conversion, intercalation and alloying electrode materials during cycling. The intercalation V$_{2}$O$_{5}$ and TiO$_{2}$ exhibit non-monotonic reversible $Λ$ and $\it M$ switching up to a factor of 1.8 ($Λ$) and 1.5 ($\it M$) as a function of lithium content. The conversion Fe$_{2}$O$_{3}$ and NiO undergo irreversible decays in $Λ$ and $\it M$ upon the first lithiation. The alloying Sb shows the largest and partially reversible order of the magnitude switching in $Λ$ between the delithiated (18 W m$^{-1}$ K$^{-1}$) and lithiated states (<1 W m$^{-1}$ K$^{-1}$). The irreversible $Λ$ is attributed to structural degradation and pulverization resulting from substantial volume changes during cycling. These findings provide new understandings of the thermal and mechanical property evolution of electrode materials during cycling of importance for battery design, and also point to pathways for forming materials with thermally switchable properties.

physics.app-ph↗