arXiv subjects

Chan Y. Park

Publications and source records attributed to Chan Y. Park.

16 recordsLinked to original sources

arXiv2026

REViT: Roto-reflection Equivariant Convolutional Vision Transformer

In this paper, we propose a discrete roto-reflection group equivariant vision transformer with convolutional attention. Roto-reflection equivariant networks preserve the rotational, flip and positional symmetry in feature maps, making them useful for tasks where orientation of the inputs is relevant to the model outputs. In image classification and object detection, most of the studies on roto-reflection equivariant models have focused on using convolutional neural networks rather than vision transformers. In this paper, we examine the challenges involved in achieving equivariance in vision transformers, and we propose a simpler way to implement a discretized roto-reflection group equivariant vision transformer. The experimental results demonstrate that our approach outperforms the existing approaches for developing discrete roto-reflection group equivariant neural networks for image classification.

cs.CV↗

arXiv2026

ErA: Error-Aware Deep Unrolling Network for Single Image Defocus Deblurring

We introduce ErA (Error-Aware Deep Unrolling Network), an end-to-end frame work for single-image defocus deblurring. ErA jointly learns a compact kerne basis and per-pixel weights, while an error-aware term in Augmented Lagrangian unrolling corrects kernel estimation errors via alternating updates and ResUNet denoisers. It achieves state-of-the-art PSNR/SSIM on DPDD, RealDOF, and RTF, and shows strong generalization on CUHK without ground truth.

eess.IV↗

arXiv2026

Drift-Augmented Scoring: Text-Derived Noise Robustness for Zero-Shot Audio-Language Classification

Contrastive audio-language models such as CLAP enable zero-shot audio classification: a sound is labelled by matching its embedding to text prompt embeddings, with no labelled audio. This matching breaks down under acoustic noise, where accuracy and mAP fall by 12-30 percentage points at 0 dB SNR on standard benchmarks. We propose Drift Augmented Scoring (DAS), a small per-class bonus added to the cosine score. The bonus rewards a class when the noisy audio embedding drifts in the direction that the class's noise-conditioned text prompts predict. It is derived from text alone, computed once and cached, and adds a single inner product per class at inference, with no gradients and no test-time batch. On a LAION CLAP backbone, we compare DAS against the four variants of Acevedo et al.'s concurrent method on UrbanSound8K and the full FSD50K eval set, mixing each clip with urban acoustic scene noise across a range of SNRs. DAS improves the metric on every test condition: by +2.60 to +5.75 accuracy points on UrbanSound8K and +1.50 to +1.74 mAP points on FSD50K.

cs.SD↗

arXiv2026

Data-Efficient Electromagnetic Surrogate Solver Through Dissipative Relaxation Transfer Learning

In neural network surrogate solvers for electromagnetic simulations, accurately modeling resonant phenomena remains a central challenge. High-amplitude resonances generate strongly localized field patterns that deviate significantly from the general distribution of non-resonant cases, leading to instability and degraded predictive performance. To address this, we introduce dissipative relaxation transfer learning (DIRTL), a data-efficient training framework that integrates transfer learning with loss-regularized optimization principles from high-Q photonics. DIRTL first pretrains the model on data generated with a small fictitious material loss, which broadens sharp resonant modes and suppresses extreme field amplitudes. This smoothing of the response landscape enables the model to learn global modal features more effectively. The pretrained model is subsequently fine-tuned on the target lossless dataset containing true high-amplitude resonances, allowing stable adaptation based on the pretrained representation. Applied to both the Fourier Neural Operator (FNO) and UNet architectures, DIRTL yields substantial improvements in prediction accuracy, including up to a two-fold error reduction for the FNO variant. Furthermore, DIRTL demonstrates robustness across diverse training conditions and supports multi-tasking performance, suggesting the generalizability and flexibility of the pretrained core. Altogether, these results position DIRTL as a physically grounded curriculum for improving the reliability of neural network surrogate solvers.

physics.optics↗

arXiv2024

Deep Joint Unrolling for Deblurring and Low-Light Image Enhancement (JUDE)

Low-light and blurring issues are prevalent when capturing photos at night, often due to the use of long exposure to address dim environments. Addressing these joint problems can be challenging and error-prone if an end-to-end model is trained without incorporating an appropriate physical model. In this paper, we introduce JUDE, a Deep Joint Unrolling for Deblurring and Low-Light Image Enhancement, inspired by the image physical model. Based on Retinex theory and the blurring model, the low-light blurry input is iteratively deblurred and decomposed, producing sharp low-light reflectance and illuminance through an unrolling mechanism. Additionally, we incorporate various modules to estimate the initial blur kernel, enhance brightness, and eliminate noise in the final image. Comprehensive experiments on LOL-Blur and Real-LOL-Blur demonstrate that our method outperforms existing techniques both quantitatively and qualitatively.

cs.CV↗

arXiv2024

Meent: Differentiable Electromagnetic Simulator for Machine Learning

Electromagnetic (EM) simulation plays a crucial role in analyzing and designing devices with sub-wavelength scale structures such as solar cells, semiconductor devices, image sensors, future displays and integrated photonic devices. Specifically, optics problems such as estimating semiconductor device structures and designing nanophotonic devices provide intriguing research topics with far-reaching real world impact. Traditional algorithms for such tasks require iteratively refining parameters through simulations, which often yield sub-optimal results due to the high computational cost of both the algorithms and EM simulations. Machine learning (ML) emerged as a promising candidate to mitigate these challenges, and optics research community has increasingly adopted ML algorithms to obtain results surpassing classical methods across various tasks. To foster a synergistic collaboration between the optics and ML communities, it is essential to have an EM simulation software that is user-friendly for both research communities. To this end, we present Meent, an EM simulation software that employs rigorous coupled-wave analysis (RCWA). Developed in Python and equipped with automatic differentiation (AD) capabilities, Meent serves as a versatile platform for integrating ML into optics research and vice versa. To demonstrate its utility as a research platform, we present three applications of Meent: 1) generating a dataset for training neural operator, 2) serving as an environment for the reinforcement learning of nanophotonic device optimization, and 3) providing a solution for inverse problems with gradient-based optimizers. These applications highlight Meent's potential to advance both EM simulation and ML methodologies. The code is available at https://github.com/kc-ml2/meent with the MIT license to promote the cross-polinations of ideas among academic researchers and industry practitioners.

cs.LG↗

arXiv2023

Physics-informed reinforcement learning for sample-efficient optimization of freeform nanophotonic devices

In the field of optics, precise control of light with arbitrary spatial resolution has long been a sought-after goal. Freeform nanophotonic devices are critical building blocks for achieving this goal, as they provide access to a design potential that could hardly be achieved by conventional fixed-shape devices. However, finding an optimal device structure in the vast combinatorial design space that scales exponentially with the number of freeform design parameters has been an enormous challenge. In this study, we propose physics-informed reinforcement learning (PIRL) as an optimization method for freeform nanophotonic devices, which combines the adjoint-based method with reinforcement learning to enhance the sample efficiency of the optimization algorithm and overcome the issue of local minima. To illustrate these advantages of PIRL over other conventional optimization algorithms, we design a family of one-dimensional metasurface beam deflectors using PIRL, obtaining more performant devices. We also explore the transfer learning capability of PIRL that further improves sample efficiency and demonstrate how the minimum feature size of the design can be enforced in PIRL through reward engineering. With its high sample efficiency, robustness, and ability to seamlessly incorporate practical device design constraints, our method offers a promising approach to highly combinatorial freeform device optimization in various physical domains.

physics.comp-ph↗

arXiv2022

NTIRE 2022 Challenge on High Dynamic Range Imaging: Methods and Results

This paper reviews the challenge on constrained high dynamic range (HDR) imaging that was part of the New Trends in Image Restoration and Enhancement (NTIRE) workshop, held in conjunction with CVPR 2022. This manuscript focuses on the competition set-up, datasets, the proposed methods and their results. The challenge aims at estimating an HDR image from multiple respective low dynamic range (LDR) observations, which might suffer from under- or over-exposed regions and different sources of noise. The challenge is composed of two tracks with an emphasis on fidelity and complexity constraints: In Track 1, participants are asked to optimize objective fidelity scores while imposing a low-complexity constraint (i.e. solutions can not exceed a given number of operations). In Track 2, participants are asked to minimize the complexity of their solutions while imposing a constraint on fidelity scores (i.e. solutions are required to obtain a higher fidelity score than the prescribed baseline). Both tracks use the same data and metrics: Fidelity is measured by means of PSNR with respect to a ground-truth HDR image (computed both directly and with a canonical tonemapping operation), while complexity metrics include the number of Multiply-Accumulate (MAC) operations and runtime (in seconds).

cs.CV↗

arXiv2021

GMAC: A Distributional Perspective on Actor-Critic Framework

In this paper, we devise a distributional framework on actor-critic as a solution to distributional instability, action type restriction, and conflation between samples and statistics. We propose a new method that minimizes the Cramér distance with the multi-step Bellman target distribution generated from a novel Sample-Replacement algorithm denoted SR($λ$), which learns the correct value distribution under multiple Bellman operations. Parameterizing a value distribution with Gaussian Mixture Model further improves the efficiency and the performance of the method, which we name GMAC. We empirically show that GMAC captures the correct representation of value distributions and improves the performance of a conventional actor-critic method with low computational cost, in both discrete and continuous action spaces using Arcade Learning Environment (ALE) and PyBullet environment.

cs.LG↗

arXiv2017

BPS Graphs: From Spectral Networks to BPS Quivers

We define "BPS graphs" on punctured Riemann surfaces associated with $A_{N-1}$ theories of class $\mathcal{S}$. BPS graphs provide a bridge between two powerful frameworks for studying the spectrum of BPS states: spectral networks and BPS quivers. They arise from degenerate spectral networks at maximal intersections of walls of marginal stability on the Coulomb branch. While the BPS spectrum is ill-defined at such intersections, a BPS graph captures a useful basis of elementary BPS states. The topology of a BPS graph encodes a BPS quiver, even for higher-rank theories and for theories with certain partial punctures. BPS graphs lead to a geometric realization of the combinatorics of Fock-Goncharov $N$-triangulations and generalize them in several ways.

hep-th↗

arXiv2017

ADE Spectral Networks and Decoupling Limits of Surface Defects

We study vacua and BPS spectra of canonical surface defects of class $\mathcal{S}$ theories in different decoupling limits using ADE spectral networks. In some regions of the IR moduli spaces of these 2d-4d systems, the mixing between 2d and 4d BPS states is suppressed, and the spectrum of 2d-4d BPS states becomes that of a 2d $\mathcal{N}=(2,2)$ theory. For some decoupling limits, we identify the 2d theories describing the surface defects with nonlinear sigma models and coset models that have been previously studied. We also study certain cases where the decoupling limit of a surface defect exhibits a set of vacua and a BPS spectrum that appear to be entirely new. A detailed analysis of these spectra and their wall-crossing behavior is performed.

hep-th↗

arXiv2016

ADE Spectral Networks

We introduce a new perspective and a generalization of spectral networks for 4d $\mathcal{N}=2$ theories of class $\mathcal{S}$ associated to Lie algebras $\mathfrak{g} = \textrm{A}_n$, $\textrm{D}_n$, $\textrm{E}_{6}$, and $\textrm{E}_{7}$. Spectral networks directly compute the BPS spectra of 2d theories on surface defects coupled to the 4d theories. A Lie algebraic interpretation of these spectra emerges naturally from our construction, leading to a new description of 2d-4d wall-crossing phenomena. Our construction also provides an efficient framework for the study of BPS spectra of the 4d theories. In addition, we consider novel types of surface defects associated with minuscule representations of $\mathfrak{g}$.

hep-th↗

arXiv2014

2d SCFT from M-branes and its spectral network

We consider the low-energy limit of the two-dimensional theory on multiple M2-branes suspended between a flat M5-brane and a curved M5-brane. We argue that it is described by an $\mathcal{N}=(2,2)$ supersymmetric Landau-Ginzburg model with the superpotential determined by the shape of the curved M5-branes, which flows in the low-energy limit to a Kazama-Suzuki coset model. We provide evidence by studying ground states and BPS spectra of the systems.

hep-th↗

arXiv2013

BPS spectrum of Argyres-Douglas theory via spectral network

We study the BPS spectrum of four-dimensional $\mathcal{N}=2$ superconformal field theory of Argyres-Douglas type, obtained via twisted compactification of six-dimensional $A_{N-1}$ $(2,0)$ theory on a sphere with an irregular puncture, by using spectral networks. We give strong evidence of the equivalence of $\mathcal{N}=2$ superconformal field theories from six-dimensional theories of different ranks by systematically comparing the chamber structure and wall-crossing phenomena.

hep-th↗

arXiv2013

2d SCFTs from M2-branes

We consider the low-energy limit of the two-dimensional theory on k M2-branes suspended between a straight M5-brane and a curved M5-brane. We argue that it is described by an N=(2,2) supersymmetric gauge theory with no matter fields but with a non-trivial twisted superpotential, and also by an N=(2,2) supersymmetric Landau-Ginzburg model, such that the (twisted) superpotentials are determined by the shape of the M5-branes. We find particular cases realize Kazama-Suzuki models. Evidence is provided by the study of ground states, chiral rings, BPS spectra and S^2 partition functions of the systems.

hep-th↗

arXiv2011

Ramification Points of Seiberg-Witten Curves

When the Seiberg-Witten curve of a four-dimensional $\mathcal{N}=2$ supersymmetric gauge theory wraps a Riemann surface as a multi-sheeted cover, a topological constraint requires that in general the curve should develop ramification points. We show that, while some of the branch points of the covering map can be identified with the punctures that appear in the work of Gaiotto, the ramification points give us additional branch points whose locations on the Riemann surface can have dependence not only on gauge coupling parameters but on Coulomb branch parameters and mass parameters of the theory. We describe how these branch points can help us to understand interesting physics in various limits of the parameters, including Argyres-Seiberg duality and Argyres-Douglas fixed points.

hep-th↗