arXiv subjects

Sanghyun Lee

Publications and source records attributed to Sanghyun Lee.

At least 37 records · Page 2Linked to original sources

arXiv2025

Where and How to Perturb: On the Design of Perturbation Guidance in Diffusion and Flow Models

Recent guidance methods in diffusion models steer reverse sampling by perturbing the model to construct an implicit weak model and guide generation away from it. Among these approaches, attention perturbation has demonstrated strong empirical performance in unconditional scenarios where classifier-free guidance is not applicable. However, existing attention perturbation methods lack principled approaches for determining where perturbations should be applied, particularly in Diffusion Transformer (DiT) architectures where quality-relevant computations are distributed across layers. In this paper, we investigate the granularity of attention perturbations, ranging from the layer level down to individual attention heads, and discover that specific heads govern distinct visual concepts such as structure, style, and texture quality. Building on this insight, we propose "HeadHunter", a systematic framework for iteratively selecting attention heads that align with user-centric objectives, enabling fine-grained control over generation quality and visual attributes. In addition, we introduce SoftPAG, which linearly interpolates each selected head's attention map toward an identity matrix, providing a continuous knob to tune perturbation strength and suppress artifacts. Our approach not only mitigates the oversmoothing issues of existing layer-level perturbation but also enables targeted manipulation of specific visual styles through compositional head selection. We validate our method on modern large-scale DiT-based text-to-image models including Stable Diffusion 3 and FLUX.1, demonstrating superior performance in both general quality enhancement and style-specific guidance. Our work provides the first head-level analysis of attention perturbation in diffusion models, uncovering interpretable specialization within attention layers and enabling practical design of effective perturbation strategies.

cs.CV↗

arXiv2025

Observation of the crossover from quantum fluxoid to half-quantum fluxoid in a chiral superconducting device

Topological superconductors are one of the intriguing material groups from the viewpoint of not only condensed matter physics but also industrial application such as quantum computers based on Majorana fermion. For the real application, developments of the thin-film topological superconductors are highly desirable. Bi/Ni bilayer is a possible candidate for thin-film chiral superconductors where the time-reversal symmetry is broken. Here we report the phase shift of resistance oscillations by half flux quantum in a ring-shaped device of epitaxial Bi/Ni bilayer induced by a small magnetic field through the ring. The half quantum fluxoid can be a decisive evidence for unconventional superconductors where the superconducting order parameter has an internal degree of freedom. The present result provides a functional operating principle for quantum devices where the phase of the supercurrent can be shifted by πwith a small magnetic field, based on the internal degree of freedom possessed by topological superconductivity.

cond-mat.supr-con↗

arXiv2025

Ensemble Score Filter for Data Assimilation of Two-Phase Flow Models in Porous Media

Numerical modeling and simulation of two-phase flow in porous media is challenging due to the uncertainties in key parameters, such as permeability. To address these challenges, we propose a computational framework by utilizing the novel Ensemble Score Filter (EnSF) to enhance the accuracy of state estimation for two-phase flow systems in porous media. The forward simulation of the two-phase flow model is implemented using a mixed finite element method, which ensures accurate approximation of the pressure, the velocity, and the saturation. The EnSF leverages score-based diffusion models to approximate filtering distributions efficiently, avoiding the computational expense of neural network-based methods. By incorporating a closed-form score approximation and an analytical update mechanism, the EnSF overcomes degeneracy issues and handles high-dimensional nonlinear filtering with minimal computational overhead. Numerical experiments demonstrate the capabilities of EnSF in scenarios with uncertain permeability and incomplete observational data.

math.NA↗

arXiv2024

A Noise is Worth Diffusion Guidance

Diffusion models excel in generating high-quality images. However, current diffusion models struggle to produce reliable images without guidance methods, such as classifier-free guidance (CFG). Are guidance methods truly necessary? Observing that noise obtained via diffusion inversion can reconstruct high-quality images without guidance, we focus on the initial noise of the denoising pipeline. By mapping Gaussian noise to `guidance-free noise', we uncover that small low-magnitude low-frequency components significantly enhance the denoising process, removing the need for guidance and thus improving both inference throughput and memory. Expanding on this, we propose \ours, a novel method that replaces guidance methods with a single refinement of the initial noise. This refined noise enables high-quality image generation without guidance, within the same diffusion pipeline. Our noise-refining model leverages efficient noise-space learning, achieving rapid convergence and strong performance with just 50K text-image pairs. We validate its effectiveness across diverse metrics and analyze how refined noise can eliminate the need for guidance. See our project page: https://cvlab-kaist.github.io/NoiseRefine/.

cs.CV↗

arXiv2024

Bound-preserving and entropy stable enriched Galerkin methods for nonlinear hyperbolic equations

In this paper, we develop monolithic limiting techniques for enforcing nonlinear stability constraints in enriched Galerkin (EG) discretizations of nonlinear scalar hyperbolic equations. To achieve local mass conservation and gain control over the cell averages, the space of continuous (multi-)linear finite element approximations is enriched with piecewise-constant functions. The resulting spatial semi-discretization has the structure of a variational multiscale method. For linear advection equations, it is inherently stable but generally not bound preserving. To satisfy discrete maximum principles and ensure entropy stability in the nonlinear case, we use limiters adapted to the structure of our locally conservative EG method. The cell averages are constrained using a flux limiter, while the nodal values of the continuous component are constrained using a clip-and-scale limiting strategy for antidiffusive element contributions. The design and analysis of our new algorithms build on recent advances in the fields of convex limiting and algebraic entropy fixes for finite element methods. In addition to proving the claimed properties of the proposed approach, we conduct numerical studies for two-dimensional nonlinear hyperbolic problems. The numerical results demonstrate the ability of our limiters to prevent violations of the imposed constraints, while preserving the optimal order of accuracy in experiments with smooth solutions.

math.NA↗

arXiv2024

A thermo-flow-mechanics-fracture model coupling a phase-field interface approach and thermo-fluid-structure interaction

Geothermal energy, a promising renewable source, relies on efficiently utilizing geothermal reservoirs, especially in Enhanced Geothermal Systems (EGS), where fractures in hot rock formations enhance permeability. Understanding fracture behavior, influenced by temperature changes, is crucial for optimizing energy extraction. To address this, we propose a novel high-accuracy phase-field interface model integrating temperature dynamics into a comprehensive hydraulic-mechanical approach, aiming for a thermo-fluid-structure interaction representation. Therein, the key technical development is a four-step algorithm. This consists of computing the fracture width, reconstructing the sharp interface geometry, solving the thermo-fluid-structure interaction (TFSI) problem, and employing a phase-field approach coupled to the temperature and pressure from the TFSI problem. By coupling temperature-hydraulic-mechanical processes with our newly proposed high-accuracy phase-field interface approach, we investigate how temperature impacts fracture width values, which are crucial for permeability in EGS reservoirs. Through this model and three different numerical simulations, we aim to provide an approach to deepen understanding of the complex interplay between temperature, mechanical deformation, and permeability evolution. Therein, we substantiate our formulations and algorithms through mesh convergence results of crack width and total crack volumes for static fractures, and crack lengths in the case of propagating fractures.

math.NA↗

arXiv2024

Continual Traffic Forecasting via Mixture of Experts

The real-world traffic networks undergo expansion through the installation of new sensors, implying that the traffic patterns continually evolve over time. Incrementally training a model on the newly added sensors would make the model forget the past knowledge, i.e., catastrophic forgetting, while retraining the model on the entire network to capture these changes is highly inefficient. To address these challenges, we propose a novel Traffic Forecasting Mixture of Experts (TFMoE) for traffic forecasting under evolving networks. The main idea is to segment the traffic flow into multiple homogeneous groups, and assign an expert model responsible for a specific group. This allows each expert model to concentrate on learning and adapting to a specific set of patterns, while minimizing interference between the experts during training, thereby preventing the dilution or replacement of prior knowledge, which is a major cause of catastrophic forgetting. Through extensive experiments on a real-world long-term streaming network dataset, PEMSD3-Stream, we demonstrate the effectiveness and efficiency of TFMoE. Our results showcase superior performance and resilience in the face of catastrophic forgetting, underscoring the effectiveness of our approach in dealing with continual learning for traffic flow forecasting in long-term streaming networks.

cs.LG↗

arXiv2024

Temporal Graph Learning Recurrent Neural Network for Traffic Forecasting

Accurate traffic flow forecasting is a crucial research topic in transportation management. However, it is a challenging problem due to rapidly changing traffic conditions, high nonlinearity of traffic flow, and complex spatial and temporal correlations of road networks. Most existing studies either try to capture the spatial dependencies between roads using the same semantic graph over different time steps, or assume all sensors on the roads are equally likely to be connected regardless of the distance between them. However, we observe that the spatial dependencies between roads indeed change over time, and two distant roads are not likely to be helpful to each other when predicting the traffic flow, both of which limit the performance of existing studies. In this paper, we propose Temporal Graph Learning Recurrent Neural Network (TGLRN) to address these problems. More precisely, to effectively model the nature of time series, we leverage Recurrent Neural Networks (RNNs) to dynamically construct a graph at each time step, thereby capturing the time-evolving spatial dependencies between roads (i.e., microscopic view). Simultaneously, we provide the Adaptive Structure Information to the model, ensuring that close and consecutive sensors are considered to be more important for predicting the traffic flow (i.e., macroscopic view). Furthermore, to endow TGLRN with robustness, we introduce an edge sampling strategy when constructing the graph at each time step, which eventually leads to further improvements on the model performance. Experimental results on four commonly used real-world benchmark datasets show the effectiveness of TGLRN.

cs.LG↗

arXiv2024

Projection-based reduced order modeling of an iterative scheme for linear thermo-poroelasticity

This paper explores an iterative coupling approach to solve linear thermo-poroelasticity problems, with its application as a high-fidelity discretization utilizing finite elements during the training of projection-based reduced order models. One of the main challenges in addressing coupled multi-physics problems is the complexity and computational expenses involved. In this study, we introduce a decoupled iterative solution approach, integrated with reduced order modeling, aimed at augmenting the efficiency of the computational algorithm. The iterative coupling technique we employ builds upon the established fixed-stress splitting scheme that has been extensively investigated for Biot's poroelasticity. By leveraging solutions derived from this coupled iterative scheme, the reduced order model employs an additional Galerkin projection onto a reduced basis space formed by a small number of modes obtained through proper orthogonal decomposition. The effectiveness of the proposed algorithm is demonstrated through numerical experiments, showcasing its computational prowess.

math.NA↗

arXiv2023

On the training and generalization of deep operator networks

We present a novel training method for deep operator networks (DeepONets), one of the most popular neural network models for operators. DeepONets are constructed by two sub-networks, namely the branch and trunk networks. Typically, the two sub-networks are trained simultaneously, which amounts to solving a complex optimization problem in a high dimensional space. In addition, the nonconvex and nonlinear nature makes training very challenging. To tackle such a challenge, we propose a two-step training method that trains the trunk network first and then sequentially trains the branch network. The core mechanism is motivated by the divide-and-conquer paradigm and is the decomposition of the entire complex training task into two subtasks with reduced complexity. Therein the Gram-Schmidt orthonormalization process is introduced which significantly improves stability and generalization ability. On the theoretical side, we establish a generalization error estimate in terms of the number of training data, the width of DeepONets, and the number of input and output sensors. Numerical examples are presented to demonstrate the effectiveness of the two-step training method, including Darcy flow in heterogeneous porous media.

math.NA↗

arXiv2022

An Enriched Galerkin Method for the Stokes Equations

We present a new enriched Galerkin (EG) scheme for the Stokes equations based on piecewise linear elements for the velocity unknowns and piecewise constant elements for the pressure. The proposed EG method augments the conforming piecewise linear space for velocity by adding an additional degree of freedom which corresponds to one discontinuous linear basis function per element. Thus, the total number of degrees of freedom is significantly reduced in comparison with standard conforming, non-conforming, and discontinuous Galerkin schemes for the Stokes equation. We show the well-posedness of the new EG approach and prove that the scheme converges optimally. For the solution of the resulting large-scale indefinite linear systems we propose robust block preconditioners, yielding scalable results independent of the discretization and physical parameters. Numerical results confirm the convergence rates of the discretization and also the robustness of the linear solvers for a variety of test problems.

math.NA↗

arXiv2021

Air Stable and Layer Dependent Ferromagnetism in Atomically Thin van der Waals CrPS$_{4}$

Ferromagnetism in two-dimensional materials presents a promising platform for the development of ultrathin spintronic devices with advanced functionalities. Recently discovered ferromagnetic van der Waals crystals such as CrI$_{3}$, readily isolated two-dimensional crystals, are highly tunable through external fields or structural modifications. However, there remains a challenge because of material instability under air exposure. Here, we report the observation of an air stable and layer dependent ferromagnetic (FM) van der Waals crystal, CrPS$_{4}$, using magneto-optic Kerr effect microscopy. In contrast to the antiferromagnetic (AFM) bulk, the FM out-of-plane spin orientation is found in the monolayer crystal. Furthermore, alternating AFM and FM properties observed in even and odd layers suggest robust antiferromagnetic exchange interactions between layers. The observed ferromagnetism in these crystals remains resilient even after the air exposure of about a day, providing possibilities for the practical applications of van der Waals spintronics.

cond-mat.mes-hall↗

arXiv2021

Three-body correlations in nonlinear response of correlated quantum liquid

Behavior of quantum liquids is a fascinating topic in physics. Even in a strongly correlated case, the linear response of a given system to an external field is described by the fluctuation-dissipation relations based on the two-body correlations in the equilibrium. However, to explore nonlinear non-equilibrium behaviors of the system beyond this well-established regime, the role of higher order correlations starting from the three-body correlations must be revealed. In this work, we experimentally investigate a controllable quantum liquid realized in a Kondo-correlated quantum dot and prove the relevance of the three-body correlations in the nonlinear conductance at finite magnetic field, which validates the recent Fermi liquid theory extended to the non-equilibrium regime.

cond-mat.mes-hall↗

arXiv2021

Cryogenic GaAs high-electron-mobility-transistor amplifier for current noise measurements

We show that a cryogenic amplifier composed of a homemade GaAs high-electron-mobility transistor (HEMT) is suitable for current-noise measurements in a mesoscopic device at dilution-refrigerator temperatures. The lower noise characteristics of our homemade HEMT leads to a lower noise floor in the experimental setup and enables more efficient current-noise measurement than is available with a commercial HEMT. We present the dc transport properties of the HEMT and the gain and noise characteristics of the amplifier. With the amplifier employed for current-noise measurements in a quantum point contact, we demonstrate the high resolution of the measurement setup by comparing it with that of the conventional one using a commercial HEMT.

cond-mat.mes-hall↗

arXiv2020

Quantum oscillations with magnetic hysteresis observed in CeTe$_{3}$ thin films

We have performed magnetotransport measurements in CeTe$_{3}$ thin films down to $0.2~{\rm K}$. It is known that CeTe$_{3}$ has two magnetic transitions at $T_{\rm N1} \approx 3~{\rm K}$ and $T_{\rm N2} \approx 1~{\rm K}$. A clear Shubnikov-de-Haas (SdH) oscillation was observed at $4~{\rm K}$, demonstrating the strong two-dimensional nature in this material. Below $T_{\rm N2}$, the SdH oscillation has two frequencies, indicating that the Fermi surface could be slightly modulated due to the second magnetic transition. We also observed a magnetic hysteresis in the SdH oscillation below $T_{\rm N1}$. Especially, there is a unique spike in the magnetoresistance at $B \approx 0.6~{\rm T}$ only when the magnetic field is swept from a high enough field (more than $2~{\rm T}$) to zero field.

cond-mat.mes-hall↗

arXiv2020

Quasi-Static Anti-Plane Shear Crack Propagation in a New Class of Nonlinear Strain-Limiting Elastic Solids using Phase-Field Regularization

We present a novel constitutive model using the framework of strain-limiting theories of elasticity for an evolution of quasi-static anti-plane fracture. The classical linear elastic fracture mechanics (LEFM), with conventional linear relationship between stress and strain, has a well documented inconsistency through which it predicts a singular cracktip strain. This clearly violates the basic tenant of the theory which is a first order approximation to finite elasticity. To overcome the issue, we investigate a new class of material models which predicts uniform and bounded strain throughout the body. The nonlinear model allows the strain value to remain small even if the stress value tends to infinity, which is achieved by an implicit relationship between stress and strain. A major objective of this paper is to couple a nonlinear bulk energy with diffusive crack employing the phase-field approach. Towards that end, an iterative L-scheme is employed and the numerical model is augmented with a penalization technique to accommodate irreversibility of crack. Several numerical experiments are presented to illustrate the capability and the performance of the proposed framework We observe the naturally bounded strain in the neighborhood of the crack-tip, leading to different bulk and crack energies for fracture propagation.

cs.CE↗

arXiv2020

Choice of Interior Penalty Coefficient for Interior Penalty Discontinuous Galerkin Method for Biot's System by Employing Machine Learning

In this paper, the optimal choice of the interior penalty parameter of the discontinuous Galerkin finite element methods for both the elliptic problems and the Biot's systems are studied by utilizing the neural network and machine learning. It is crucial to choose the optimal interior penalty parameter, which is not too small or not too large for the stability, robustness, and efficiency of the numerical discretized solutions. Both linear regression and nonlinear artificial neural network methods are employed and compared using several numerical experiments to illustrate the capability of our proposed computational framework. This framework is an integral part of a developing automated numerical simulation platform because it can automatically identify the optimal interior penalty parameter. Real-time feedback could also be implemented to update and improve model accuracy on the fly.

math.NA↗

arXiv2020

Nonlinear Strain-limiting Elasticity for Fracture Propagation with Phase-Field Approach

The conventional model governing the spread of fractures in elastic material is formulated by coupling linear elasticity with deformation systems. The classical linear elastic fracture mechanics (LEFM) model is derived based on the assumption of small strain values. However, since the strain values in the model are linearly proportional to the stress values, the strain value can be large if the stress value increases. Thus this results in the contradiction of the assumption to LEFM and it is one of the major disadvantages of the model. In particular, this singular behavior of the strain values is often observed especially near the crack-tip, and it may not accurately predict realistic phenomena. Thus, we investigate the framework of a new class of theoretical model, which is known as the nonlinear strain-limiting model. The advantage of the nonlinear strain-limiting models over LEFM is that the strain value remains bounded even if the stress value tends to the infinity. This is achieved by assuming the nonlinear relation between the strain and stress in the derivation of the model. Moreover, we consider the quasi-static fracture propagation by coupling with the phase-field approach to present the effectiveness of the proposed strain-limiting model. Several numerical examples to evaluate and validate the performance of the new model and algorithms are presented. Detailed comparisons of the strain values, fracture energy, and fracture propagation speed between nonlinear strain-limiting model and LEFM for the quasi-static fracture propagation are discussed.

cs.CE↗