arXiv subjects

Shinjae Yoo

Publications and source records attributed to Shinjae Yoo.

At least 91 records · Page 5Linked to original sources

arXiv2023

UVCGAN v2: An Improved Cycle-Consistent GAN for Unpaired Image-to-Image Translation

An unpaired image-to-image (I2I) translation technique seeks to find a mapping between two domains of data in a fully unsupervised manner. While initial solutions to the I2I problem were provided by generative adversarial neural networks (GANs), diffusion models (DMs) currently hold the state-of-the-art status on the I2I translation benchmarks in terms of Frechet inception distance (FID). Yet, DMs suffer from limitations, such as not using data from the source domain during the training or maintaining consistency of the source and translated images only via simple pixel-wise errors. This work improves a recent UVCGAN model and equips it with modern advancements in model architectures and training procedures. The resulting revised model significantly outperforms other advanced GAN- and DM-based competitors on a variety of benchmarks. In the case of Male-to-Female translation of CelebA, the model achieves more than 40% improvement in FID score compared to the state-of-the-art results. This work also demonstrates the ineffectiveness of the pixel-wise I2I translation faithfulness metrics and suggests their revision. The code and trained models are available at https://github.com/LS4GAN/uvcgan2

cs.CV↗

arXiv2023

Transferable Graph Neural Fingerprint Models for Quick Response to Future Bio-Threats

Fast screening of drug molecules based on the ligand binding affinity is an important step in the drug discovery pipeline. Graph neural fingerprint is a promising method for developing molecular docking surrogates with high throughput and great fidelity. In this study, we built a COVID-19 drug docking dataset of about 300,000 drug candidates on 23 coronavirus protein targets. With this dataset, we trained graph neural fingerprint docking models for high-throughput virtual COVID-19 drug screening. The graph neural fingerprint models yield high prediction accuracy on docking scores with the mean squared error lower than $0.21$ kcal/mol for most of the docking targets, showing significant improvement over conventional circular fingerprint methods. To make the neural fingerprints transferable for unknown targets, we also propose a transferable graph neural fingerprint method trained on multiple targets. With comparable accuracy to target-specific graph neural fingerprint models, the transferable model exhibits superb training and data efficiency. We highlight that the impact of this study extends beyond COVID-19 dataset, as our approach for fast virtual ligand screening can be easily adapted and integrated into a general machine learning-accelerated pipeline to battle future bio-threats.

q-bio.BM↗

arXiv2023

Exploring Robust Features for Improving Adversarial Robustness

While deep neural networks (DNNs) have revolutionized many fields, their fragility to carefully designed adversarial attacks impedes the usage of DNNs in safety-critical applications. In this paper, we strive to explore the robust features which are not affected by the adversarial perturbations, i.e., invariant to the clean image and its adversarial examples, to improve the model's adversarial robustness. Specifically, we propose a feature disentanglement model to segregate the robust features from non-robust features and domain specific features. The extensive experiments on four widely used datasets with different attacks demonstrate that robust features obtained from our model improve the model's adversarial robustness compared to the state-of-the-art approaches. Moreover, the trained domain discriminator is able to identify the domain specific features from the clean images and adversarial examples almost perfectly. This enables adversarial example detection without incurring additional computational costs. With that, we can also specify different classifiers for clean images and adversarial examples, thereby avoiding any drop in clean image accuracy.

cs.CV↗

arXiv2023

INSURE: An Information Theory Inspired Disentanglement and Purification Model for Domain Generalization

Domain Generalization (DG) aims to learn a generalizable model on the unseen target domain by only training on the multiple observed source domains. Although a variety of DG methods have focused on extracting domain-invariant features, the domain-specific class-relevant features have attracted attention and been argued to benefit generalization to the unseen target domain. To take into account the class-relevant domain-specific information, in this paper we propose an Information theory iNspired diSentanglement and pURification modEl (INSURE) to explicitly disentangle the latent features to obtain sufficient and compact (necessary) class-relevant feature for generalization to the unseen domain. Specifically, we first propose an information theory inspired loss function to ensure the disentangled class-relevant features contain sufficient class label information and the other disentangled auxiliary feature has sufficient domain information. We further propose a paired purification loss function to let the auxiliary feature discard all the class-relevant information and thus the class-relevant feature will contain sufficient and compact (necessary) class-relevant information. Moreover, instead of using multiple encoders, we propose to use a learnable binary mask as our disentangler to make the disentanglement more efficient and make the disentangled features complementary to each other. We conduct extensive experiments on four widely used DG benchmark datasets including PACS, OfficeHome, TerraIncognita, and DomainNet. The proposed INSURE outperforms the state-of-art methods. We also empirically show that domain-specific class-relevant features are beneficial for domain generalization.

cs.CV↗

arXiv2023

Neural Network Based Point Spread Function Deconvolution For Astronomical Applications

Optical astronomical images are strongly affected by the point spread function (PSF) of the optical system and the atmosphere (seeing) which blurs the observed image. The amount of blurring depends both on the observed band, and on the atmospheric conditions during observation. A typical astronomical image will likely have a unique PSF, that is non-circular and different in different bands. At the same time, observations of known stars also give us an accurate determination of this PSF. Therefore, any serious candidate for production analysis of astronomical images must take the known PSF into account during the image analysis. So far, the majority of applications of neural networks (NN) to astronomical image analysis have ignored this problem by assuming a fixed PSF in training and validation. We present a neural-network based deconvolution algorithm based on Deep Wiener Deconvolution Network (DWDN). This algorithm belongs to a class of non-blind deconvolution algorithms, since it assumes the PSF shape is known. We study the performance of different versions of this algorithm under realistic observational conditions in terms of the recovery of the most relevant astronomical quantities such as colors, ellipticities and orientations. We investigate custom loss functions that optimize the recovery of astronomical quantities with mixed results.

astro-ph.IM↗

arXiv2023

An extensible point-based method for data chart value detection

We present an extensible method for identifying semantic points to reverse engineer (i.e. extract the values of) data charts, particularly those in scientific articles. Our method uses a point proposal network (akin to region proposal networks for object detection) to directly predict the position of points of interest in a chart, and it is readily extensible to multiple chart types and chart elements. We focus on complex bar charts in the scientific literature, on which our model is able to detect salient points with an accuracy of 0.8705 F1 (@1.5-cell max deviation); it achieves 0.9810 F1 on synthetically-generated charts similar to those used in prior works. We also explore training exclusively on synthetic data with novel augmentations, reaching surprisingly competent performance in this way (0.6621 F1) on real charts with widely varying appearance, and we further demonstrate our unchanged method applied directly to synthetic pie charts (0.8343 F1). Datasets, trained models, and evaluation code are available at https://github.com/BNLNLP/PPN_model.

cs.CV↗

arXiv2023

Emulating Expert Insight: A Robust Strategy for Optimal Experimental Design

The challenge of optimal design of experiments (DOE) pervades materials science, physics, chemistry, and biology. Bayesian optimization has been used to address this challenge in vast sample spaces, although it requires framing experimental campaigns through the lens of maximizing some observable. This framing is insufficient for epistemic research goals that seek to comprehensively analyze a sample space, without an explicit scalar objective (e.g., the characterization of a wafer or sample library). In this work, we propose a flexible formulation of scientific value that recasts a dataset of input conditions and higher-dimensional observable data into a continuous, scalar metric. Intuitively, the scientific value function measures where observables change significantly, emulating the perspective of experts driving an experiment, and can be used in collaborative analysis tools or as an objective for optimization techniques. We demonstrate this technique by exploring simulated phase boundaries from different observables, autonomously driving a variable temperature measurement of a ferroelectric material, and providing feedback from a nanoparticle synthesis campaign. The method is seamlessly compatible with existing optimization tools, can be extended to multi-modal and multi-fidelity experiments, and can integrate existing models of an experimental system. Because of its flexibility, it can be deployed in a range of experimental settings for autonomous or accelerated experiments.

cond-mat.mtrl-sci↗

arXiv2023

Quantum Computing for High-Energy Physics: State of the Art and Challenges. Summary of the QC4HEP Working Group

Quantum computers offer an intriguing path for a paradigmatic change of computing in the natural sciences and beyond, with the potential for achieving a so-called quantum advantage, namely a significant (in some cases exponential) speed-up of numerical simulations. The rapid development of hardware devices with various realizations of qubits enables the execution of small scale but representative applications on quantum computers. In particular, the high-energy physics community plays a pivotal role in accessing the power of quantum computing, since the field is a driving source for challenging computational problems. This concerns, on the theoretical side, the exploration of models which are very hard or even impossible to address with classical techniques and, on the experimental side, the enormous data challenge of newly emerging experiments, such as the upgrade of the Large Hadron Collider. In this roadmap paper, led by CERN, DESY and IBM, we provide the status of high-energy physics quantum computations and give examples for theoretical and experimental target benchmark applications, which can be addressed in the near future. Having the IBM 100 x 100 challenge in mind, where possible, we also provide resource estimates for the examples given using error mitigated quantum computing.

quant-ph↗

arXiv2023

Accurate, uncertainty-aware classification of molecular chemical motifs from multi-modal X-ray absorption spectroscopy

Accurate classification of molecular chemical motifs from experimental measurement is an important problem in molecular physics, chemistry and biology. In this work, we present neural network ensemble classifiers for predicting the presence (or lack thereof) of 41 different chemical motifs on small molecules from simulated C, N and O K-edge X-ray absorption near-edge structure (XANES) spectra. Our classifiers not only reach a maximum average class-balanced accuracy of 0.99 but also accurately quantify uncertainty. We also show that including multiple XANES modalities improves predictions notably on average, demonstrating a "multi-modal advantage" over any single modality. In addition to structure refinement, our approach can be generalized for broad applications with molecular design pipelines.

cond-mat.mtrl-sci↗

arXiv2023

Optimal sensor placement for reconstructing wind pressure field around buildings using compressed sensing

Deciding how to optimally deploy sensors in a large, complex, and spatially extended structure is critical to ensure that the surface pressure field is accurately captured for subsequent analysis and design. In some cases, reconstruction of missing data is required in downstream tasks such as the development of digital twins. This paper presents a data-driven sparse sensor selection algorithm, aiming to provide the most information contents for reconstructing aerodynamic characteristics of wind pressures over tall building structures parsimoniously. The algorithm first fits a set of basis functions to the training data, then applies a computationally efficient QR algorithm that ranks existing pressure sensors in order of importance based on the state reconstruction to this tailored basis. The findings of this study show that the proposed algorithm successfully reconstructs the aerodynamic characteristics of tall buildings from sparse measurement locations, generating stable and optimal solutions across a range of conditions. As a result, this study serves as a promising first step toward leveraging the success of data-driven and machine learning algorithms to supplement traditional genetic algorithms currently used in wind engineering.

physics.flu-dyn↗

arXiv2023

A Neural PDE Solver with Temporal Stencil Modeling

Numerical simulation of non-linear partial differential equations plays a crucial role in modeling physical science and engineering phenomena, such as weather, climate, and aerodynamics. Recent Machine Learning (ML) models trained on low-resolution spatio-temporal signals have shown new promises in capturing important dynamics in high-resolution signals, under the condition that the models can effectively recover the missing details. However, this study shows that significant information is often lost in the low-resolution down-sampled features. To address such issues, we propose a new approach, namely Temporal Stencil Modeling (TSM), which combines the strengths of advanced time-series sequence modeling (with the HiPPO features) and state-of-the-art neural PDE solvers (with learnable stencil modeling). TSM aims to recover the lost information from the PDE trajectories and can be regarded as a temporal generalization of classic finite volume methods such as WENO. Our experimental results show that TSM achieves the new state-of-the-art simulation accuracy for 2-D incompressible Navier-Stokes turbulent flows: it significantly outperforms the previously reported best results by 19.9% in terms of the highly-correlated duration time and reduces the inference latency into 80%. We also show a strong generalization ability of the proposed method to various out-of-distribution turbulent flow settings. Our code is available at "https://github.com/Edward-Sun/TSM-PDE".

cs.LG↗

arXiv2023

Simulated sulfur K-edge X-ray absorption spectroscopy database of lithium thiophosphate solid electrolytes

X-ray absorption spectroscopy (XAS) is a premier technique for materials characterization, providing key information about the local chemical environment of the absorber atom. In this work, we develop a database of sulfur K-edge XAS spectra of crystalline and amorphous lithium thiophosphate materials based on the atomic structures reported in Chem. Mater., 34, 6702 (2022). The XAS database is based on simulations using the excited electron and core-hole pseudopotential approach implemented in the Vienna Ab initio Simulation Package. Our database contains 2681 S K-edge XAS spectra for 66 crystalline and glassy structure models, making it the largest collection of first-principles computational XAS spectra for glass/ceramic lithium thiophosphates to date. This database can be used to correlate S spectral features with distinct S species based on their local coordination and short-range ordering in sulfide-based solid electrolytes. The data is openly distributed via the Materials Cloud, allowing researchers to access it for free and use it for further analysis, such as spectral fingerprinting, matching with experiments, and developing machine learning models.

cond-mat.mtrl-sci↗

arXiv2023

Uncertainty-aware predictions of molecular X-ray absorption spectra using neural network ensembles

As machine learning (ML) methods continue to be applied to a broad scope of problems in the physical sciences, uncertainty quantification is becoming correspondingly more important for their robust application. Uncertainty aware machine learning methods have been used in select applications, but largely for scalar properties. In this work, we showcase an exemplary study in which neural network ensembles are used to predict the X-ray absorption spectra of small molecules, as well as their point-wise uncertainty, from local atomic environments. The performance of the resulting surrogate clearly demonstrates quantitative correlation between errors relative to ground truth and the predicted uncertainty estimates. Significantly, the model provides an upper bound on the expected error. Specifically, an important quality of this uncertainty-aware model is that it can indicate when the model is predicting on out-of-sample data. This allows for its integration with large scale sampling of structures together with active learning or other techniques for structure refinement. Additionally, our models can be generalized to larger molecules than those used for training, and also successfully track uncertainty due to random distortions in test molecules. While we demonstrate this workflow on a specific example, ensemble learning is completely general. We believe it could have significant impact on ML-enabled forward modeling of a broad array of molecular and materials properties.

cond-mat.mtrl-sci↗

arXiv2023

Comprehensive analysis of gene expression profiles to radiation exposure reveals molecular signatures of low-dose radiation response

There are various sources of ionizing radiation exposure, where medical exposure for radiation therapy or diagnosis is the most common human-made source. Understanding how gene expression is modulated after ionizing radiation exposure and investigating the presence of any dose-dependent gene expression patterns have broad implications for health risks from radiotherapy, medical radiation diagnostic procedures, as well as other environmental exposure. In this paper, we perform a comprehensive pathway-based analysis of gene expression profiles in response to low-dose radiation exposure, in order to examine the potential mechanism of gene regulation underlying such responses. To accomplish this goal, we employ a statistical framework to determine whether a specific group of genes belonging to a known pathway display coordinated expression patterns that are modulated in a manner consistent with the radiation level. Findings in our study suggest that there exist complex yet consistent signatures that reflect the molecular response to radiation exposure, which differ between low-dose and high-dose radiation.

q-bio.GN↗

arXiv2022

UVCGAN: UNet Vision Transformer cycle-consistent GAN for unpaired image-to-image translation

Unpaired image-to-image translation has broad applications in art, design, and scientific simulations. One early breakthrough was CycleGAN that emphasizes one-to-one mappings between two unpaired image domains via generative-adversarial networks (GAN) coupled with the cycle-consistency constraint, while more recent works promote one-to-many mapping to boost diversity of the translated images. Motivated by scientific simulation and one-to-one needs, this work revisits the classic CycleGAN framework and boosts its performance to outperform more contemporary models without relaxing the cycle-consistency constraint. To achieve this, we equip the generator with a Vision Transformer (ViT) and employ necessary training and regularization techniques. Compared to previous best-performing models, our model performs better and retains a strong correlation between the original and translated image. An accompanying ablation study shows that both the gradient penalty and self-supervised pre-training are crucial to the improvement. To promote reproducibility and open science, the source code, hyperparameter configurations, and pre-trained model are available at https://github.com/LS4GAN/uvcgan.

cs.CV↗

arXiv2022

Galaxy Deblending using Residual Dense Neural networks

We present a new neural network approach for deblending galaxy images in astronomical data using Residual Dense Neural network (RDN) architecture. We train the network on synthetic galaxy images similar to the typical arrangements of field galaxies with a finite point spread function (PSF) and realistic noise levels. The main novelty of our approach is the usage of two distinct neural networks: i) a deblending network which isolates a single galaxy postage stamp from the composite and, ii) a classifier network which counts the remaining number of galaxies. The deblending proceeds by iteratively peeling one galaxy at a time from the composite until the image contains no further objects as determined by the classifier, or by other stopping criteria. By looking at the consistency in the outputs of the two networks, we can assess the quality of the deblending. We characterize the flux and shape reconstructions in different quality bins and compare our deblender with the industry standard, SExtractor. We also discuss possible future extensions for the project with variable PSFs and noise levels.

astro-ph.GA↗

arXiv2022

DUNE Software and High Performance Computing

DUNE, like other HEP experiments, faces a challenge related to matching execution patterns of our production simulation and data processing software to the limitations imposed by modern high-performance computing facilities. In order to efficiently exploit these new architectures, particularly those with high CPU core counts and GPU accelerators, our existing software execution models require adaptation. In addition, the large size of individual units of raw data from the far detector modules pose an additional challenge somewhat unique to DUNE. Here we describe some of these problems and how we begin to solve them today with existing software frameworks and toolkits. We also describe ways we may leverage these existing software architectures to attack remaining problems going forward. This whitepaper is a contribution to the Computational Frontier of Snowmass21.

hep-ex↗

arXiv2022

A Deep Finite Difference Emulator for the Fast Simulation of Coupled Viscous Burgers' Equation

This work proposes a deep learning-based emulator for the efficient computation of the coupled viscous Burgers' equation with random initial conditions. In a departure from traditional data-driven deep learning approaches, the proposed emulator does not require a classical numerical solver to collect training data. Instead, it makes direct use of the problem's physics. Specifically, the model emulates a second-order finite difference solver, i.e., the Crank-Nicolson scheme in learning dynamics. A systematic case study is conducted to examine the model's prediction performance, generalization ability, and computational efficiency. The computed results are graphically represented and compared to those of state-of-the-art numerical solvers.

physics.comp-ph↗