arXiv subjects

Joon-Hyuk Ko

Publications and source records attributed to Joon-Hyuk Ko.

6 recordsLinked to original sources

arXiv2026

Watch your neighbors: Training statistically accurate chaotic systems with local phase space information

Chaotic systems pose fundamental challenges for data-driven dynamics discovery, as small modeling errors lead to exponentially growing trajectory discrepancies. Since exact long-term prediction is unattainable, it is natural to ask what a good surrogate model for chaotic dynamics is. Prior work has largely focused either on reproducing the Jacobian of the underlying dynamics, which governs local expansion and contraction rates, or on training surrogate models that reproduce the ground-truth dynamics' long-term statistical behavior. In this work, we propose a new framework that aims to bridge these two paradigms by training surrogate dynamics models with accurate Jacobians and long-term statistical properties. Our method constructs a local covering of a chaotic attractor in phase space and analyzes the expansion and contraction of these coverings under the dynamics. The surrogate model is trained by minimizing the maximum mean discrepancy between the pushforward distributions of the coverings under the surrogate and ground-truth dynamics. Experiments show that our method significantly improves Jacobian accuracy while remaining competitive with state-of-the-art statistically accurate dynamics learning methods. Our code is fully available at https://anonymous.4open.science/r/neighborwatch.

cs.LG↗

arXiv2026

A Robust In-Context Model for Conservation Laws: Injecting Context into Flux Neural Operators via Recurrent Vision Transformers

We propose an architecture that augments the Flux Neural Operator (Flux NO), which combines the classical finite volume method (FVM) with neural operators, with ViT-based context injection. Our model is formulated as a hypernetwork: it extracts solution dynamics over a finite temporal window, encodes them with a recurrent Vision Transformer, and generates the parameters of a context-conditioned neural operator. This enables the model to infer and solve conservation laws without explicit access to the governing equation or PDE coefficients. Experimentally, we show that the proposed method preserves the robustness, generalization ability, and long-time prediction advantages of Flux NO over standard neural operators, while delivering reliable numerical solutions across a broad range of conservative systems, including previously unseen fluxes. Our code is available at https://github.com/xx257xx/CONTEXT_FLUX_NO.

cs.LG↗

arXiv2026

A comprehensive semi-automated fabrication system for quartz tuning fork AFM probe with real-time resonance frequency monitoring and Q-factor control

Quartz tuning fork-based atomic force microscopy (QTF-AFM) has become a powerful tool for high-resolution imaging of both conductive and insulating samples, including semiconductor structures and metal-coated surfaces as well as soft matter under ambient conditions, while also enabling measurements in more demanding environments including ultrahigh vacuum and cryogenic conditions where conventional cantilever-based AFM often encounters limitations. However, the broader adoption of QTF-AFM has been constrained by the difficulty of attaching a cantilever tip to a quartz tuning fork (QTF) with the positional and angular precision required for repeatable and reproducible probe fabrication. For stable operation, the tip must be placed precisely at the midline of a single tine, aligned parallel to the prong axis, and rigidly secured. Even slight lateral offsets or angular deviations disrupt the intrinsic antisymmetric flexural mode, induce torsional coupling, and ultimately lead to systematic image distortions and reduced measurement integrity. In this work, we present a comprehensive, semi-automated QTF-tip fabrication system that integrates precision alignment, real-time frequency-sweep monitoring, and controlled Q-factor tuning within a single workflow. Experimental characterization demonstrates consistent probe preparation across multiple trials, preservation of sharp and well-defined resonance responses with deliberately adjustable damping, and high-fidelity, high-resolution imaging in practical scanning tests. This integrated approach provides a reproducible framework to QTF-based probe fabrication, lowering the technical barrier to QTF-AFM implementation and broadening its applicability across diverse sample types and operating environments.

cond-mat.mes-hall↗

arXiv2024

Homotopy-based training of NeuralODEs for accurate dynamics discovery

Neural Ordinary Differential Equations (NeuralODEs) present an attractive way to extract dynamical laws from time series data, as they bridge neural networks with the differential equation-based modeling paradigm of the physical sciences. However, these models often display long training times and suboptimal results, especially for longer duration data. While a common strategy in the literature imposes strong constraints to the NeuralODE architecture to inherently promote stable model dynamics, such methods are ill-suited for dynamics discovery as the unknown governing equation is not guaranteed to satisfy the assumed constraints. In this paper, we develop a new training method for NeuralODEs, based on synchronization and homotopy optimization, that does not require changes to the model architecture. We show that synchronizing the model dynamics and the training data tames the originally irregular loss landscape, which homotopy optimization can then leverage to enhance training. Through benchmark experiments, we demonstrate our method achieves competitive or better training loss while often requiring less than half the number of training epochs compared to other model-agnostic techniques. Furthermore, models trained with our method display better extrapolation capabilities, highlighting the effectiveness of our method.

cs.LG↗

arXiv2020

Bayesian neural network with pretrained protein embedding enhances prediction accuracy of drug-protein interaction

The characterization of drug-protein interactions is crucial in the high-throughput screening for drug discovery. The deep learning-based approaches have attracted attention because they can predict drug-protein interactions without trial-and-error by humans. However, because data labeling requires significant resources, the available protein data size is relatively small, which consequently decreases model performance. Here we propose two methods to construct a deep learning framework that exhibits superior performance with a small labeled dataset. At first, we use transfer learning in encoding protein sequences with a pretrained model, which trains general sequence representations in an unsupervised manner. Second, we use a Bayesian neural network to make a robust model by estimating the data uncertainty. As a result, our model performs better than the previous baselines for predicting drug-protein interactions. We also show that the quantified uncertainty from the Bayesian inference is related to the confidence and can be used for screening DPI data points.

cs.LG↗

arXiv2020

GCIceNet: A Graph Convolutional Network for Accurate Classification of Water Phases

Understanding phases of water molecules based on local structure is essential for understanding their anomalous properties. However, due to complicated structural motifs formed via hydrogen bonds, conventional order parameters represent the water molecules incompletely. In this paper, we develop a GCIceNet, which automatically generates machine-based order parameters for classifying the phases of the water molecules via supervised and unsupervised learning. Multiple graph convolutional layers in the GCIceNet can learn topological informations of the complex hydrogen bond networks. It shows a substantial improvement of accuracy for predicting the phase of water molecules in the bulk system and the ice/vapor interface system. A relative importance analysis shows that the GCIceNet can capture the structural features of the given system hidden in the input data. Augmented with the vast amount of data provided by molecular dynamics simulations, the GCIceNet is expected to serve as a powerful tool for the fields of glassy liquids and hydration layers around biomolecules.

cond-mat.soft↗