Dong-Hyun Won

Prof. Dr. Dong-Hyun Won


Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
20 February 2026 in Electronics
Verified
Journal Article
Open access
PDF file
DWDong-Hyun Won
Dong-Hyun Won
KSKwang Seong Shin
Kwang Seong Shin
SYSungkwan Youm
Sungkwan Youm
Class imbalance in tabular datasets poses a challenge for machine learning classification tasks, often leading to biased models that underperform in predicting minority class instances. This study presents a comparative analysis of synthetic data generation methods for addressing class imbalance in tabular data. We evaluate four augmentation approaches—Synthetic Minority Over-sampling Technique (SMOTE), Gaussian Copula, Tabular Variational Autoencoder (TVAE), and Conditional Tabular Generative Adversarial Network (CTGAN)—using the University of California Irvine (UCI) Bank Marketing dataset, which exhibits a class imbalance ratio of approximately 7.88:1. Our experimental framework assesses each method across three dimensions: statistical fidelity to the original data distribution evaluated through four complementary metrics (marginal numerical similarity, categorical distribution similarity, correlation structure preservation, and Kolmogorov–Smirnov test), machine learning utility measured through classification performance, and minority class detection capability. Results indicate that all augmentation methods achieved statistically significant improvements over the baseline (p<0.05). SMOTE achieved the highest recall (54.2%, a 117.6% relative improvement over the baseline) and F1-Score (0.437, +22.4% over the baseline) for minority class detection, while Gaussian Copula provided the highest composite fidelity score (0.930) with competitive predictive performance. A weak negative correlation (ρ=−0.30) between composite fidelity and classification performance was observed, suggesting that higher statistical fidelity does not necessarily translate to better downstream task performance. Deep learning-based methods (TVAE, CTGAN) showed statistically significant improvements over the baseline (recall: +58% to +63%) but underperformed compared to simpler methods under default configurations, suggesting the need for larger training samples or more extensive hyperparameter tuning. These findings offer reference points for practitioners working with moderately imbalanced tabular data with limited minority class samples, supporting the selection of generation strategies based on specific requirements regarding data fidelity and classification objectives.

4 August 2026 in Electronics
Verified
Journal Article
Open access
PDF file
DWDong-Hyun Won
Dong-Hyun Won
HPHyuk-Gyu Park
Hyuk-Gyu Park
KSKwang Seong Shin
Kwang Seong Shin
Labeling images is expensive, but unlabeled data is abundant. Semi-supervised learning (SSL) addresses this gap, though the dominant pseudo-labeling methods can suffer from confirmation bias—reinforcing their own confident-but-wrong predictions—most severely when labels are scarcest. Under a controlled, reproducible compute-constrained protocol on CIFAR-10, SVHN, and CIFAR-100, we examine which ingredients of consistency-based SSL actually help when as few as four labels per class are available. We propose CCM (Confidence-Calibrated Consistency Matching)—a per-class curriculum threshold, a dual strong-view consistency loss, and a smooth confidence weighting that softly admits borderline pseudo-labels—together with a unified view in which FixMatch, FlexMatch, SoftMatch, and CCM instantiate a single generalized weighting function. On CIFAR-10 with 40 labels, CCM reaches 35.14%, a significant improvement over the FlexMatch design it directly extends (+3.30 percentage points (pp), paired t-test p = 0.001). SoftMatch, re-trained under the identical budget, performs better still at the two smallest budgets (37.04% at 40 labels, p = 0.049), while CCM leads numerically at 4000 labels: the two smooth-weighting designs top the extreme-scarcity board—convergent evidence that the smoothness of the weighting function, more than the placement of its threshold, is the decisive design axis. We also report a negative result: cross-view agreement helps neither as an admission gate nor as reliability reweighting, reducing accuracy by up to 4.37 pp; agreement is a positive correctness signal, but its absolute level (approximately 44% correct among agreeing pseudo-labels) is too low to filter on safely. Curriculum thresholding, by contrast, hurts on the easier SVHN dataset and fails outright on CIFAR-100 when its per-class statistics become too thin. CCM adds no inference-time cost. We do not claim universality; we characterize when each ingredient helps within a single, identical-budget protocol.

6 August 2026 in Electronics
Verified
Journal Article
Open access
PDF file
DWDong-Hyun Won
Dong-Hyun Won
KSKwang Seong Shin
Kwang Seong Shin
Background: Distribution shift is one of the most pervasive data-related challenges for time-series classification. In sensor-based Human Activity Recognition (HAR), a model trained on one population of subjects often degrades substantially when deployed on another, even with identical hardware and activity labels. This paper addresses that setting specifically: cross-subject distribution shift. Methods: We propose FTDA (Frequency–Time Domain Alignment), an unsupervised domain-adaptation method that processes raw signals and their FFT-magnitude spectra through a dual-branch encoder, combines a Gradient-Reversal adversarial loss on the joint feature with a Multi-Kernel MMD on the frequency branch, and applies a symmetric-KL time–frequency consistency loss on unlabeled target data. Results: On UCI HAR, over four cross-subject transfer tasks and 10 random seeds (n=40 paired observations), FTDA reaches 95.01% target accuracy against 92.46% for Source-only, 92.29% for DANN and 93.32% for a RAINCOAT-style baseline. Because the tasks share subjects and two of them are mirror configurations, no pooled significance test is offered; inference is confined to within-task paired tests (10 seeds each, Holm-corrected), under which FTDA is ahead of every single-view baseline on all four tasks, while the difference from the two dual-view competitors (+1.69 pp against RAINCOAT-style, descriptive task-level interval [−0.40,+3.78]) is not established. Over a common set of seeds, controls widened to FTDA’s parameter and multiply-accumulate budgets shifted accuracy by at most +0.35 pp against DANN (intervals including zero), where FTDA gains +2.97 pp on the same runs, so added capacity alone does not appear to explain the improvement, and a target-only reference model, which is optimistic because it is evaluated on the windows it was trained on, sits 0.70 pp above FTDA. An ablation shows that the frequency branch is useful chiefly because it enables a coupled mechanism of cross-view consistency and auxiliary source supervision (+2.02 pp for that final step) rather than concatenated features (+1.53 pp); for the frequency MMD term, no measurable benefit was detected (−0.57 pp, 95% CI [−1.28,+0.13]), and it can be removed. On a second dataset, HHAR, the cross-subject advantage does not reproduce under cross-device shift: we detect no difference from time-only adversarial baselines (paired difference +0.48 pp against DANN, 95% CI [−1.20,+2.16]; exploratory, since all device pairs share the same subjects), so the benefit of the spectral view does not appear to carry across shift types. Conclusions: Under cross-subject shift, the operative ingredient is not the spectral view as an extra set of features but the coupled consistency-and-auxiliary mechanism that the second view makes possible; the alignment objectives themselves contribute little, and no measurable benefit was detected for the frequency MMD term. FTDA realizes this view at a modest 0.31 M/7.55 MMAC footprint with no OT solver in the training loop and no hyper-parameter that requires target labels. Transfer across devices and operating environments is not evaluated here and remains open for future work.

27 October 2026 in Electronics
Verified
Journal Article
Open access
PDF file
HPHyuk-Gyu Park
Hyuk-Gyu Park
KSKwang Seong Shin
Kwang Seong Shin
DWDong-Hyun Won
Dong-Hyun Won
Attribute noise, the measurement error in feature values, is a data-related challenge distinct from label noise, and the noise present at training time rarely equals the noise present at deployment. This article studies this train–test noise mismatch in tabular classification. Its main contribution is noise-matched training (NMT), which estimates a feature-wise noise level from the variance excess of unlabeled deployment inputs over the training inputs, with a threshold calibrated on the training split, and injects noise of that level in one additional fit. Its identifiability assumption, that the clean deployment variance equals the training variance, is stated and tested. In a one-dimensional Gaussian model, the excess risk of a rule tuned at the wrong noise level vanishes only at the matched level. Five learners on ten public OpenML/UCI datasets were evaluated under eight test-time noise conditions and twelve training conditions, with the dataset as the unit of inference. The matched level was, on average, the best fixed training level, except for nearest-neighbor classification. NMT had the highest average accuracy and helped most when only some features were noisy, but none of its dataset-level comparisons remained significant after multiplicity adjustment. It gave no benefit under impulsive noise, for nearest neighbors, or under covariate shift without noise.

16 November 2026 in Electronics
Verified
Journal Article
Open access
PDF file
DWDong-Hyun Won
Dong-Hyun Won
KSKwang Seong Shin
Kwang Seong Shin
Loss corrections based on the noise transition matrix T promise to undo the annotation errors that deep networks memorize in medical images. On six public MedMNIST benchmarks with synthetic label noise, we separately evaluate the estimation of T, the selection of clean examples, and the clean test accuracy. Two propositions show that backward correction inflates the loss variance and that forward correction with an estimate changes decisions only where the clean margin is below twice an explicit error measure of the estimate. In 2175 runs, backward correction was below cross-entropy in 29 of 30 noisy settings, and the cost of forward correction grew with this error measure. The proposed validation-gated forward correction (VGFC), which chooses between forward correction and small-loss selection without clean labels or a known noise rate, was more accurate on average than either candidate but not clearly better than a disagreement-gated variant (+0.13 percentage points, 95% CI [−1.41, +1.24]). Post hoc, it was close to self-adaptive training on the four datasets with accurate estimates of T (above it with ResNet-18 on three of them) and below it on the other two. The benefit of transition-based methods is thus bounded by the accuracy of the estimated matrix. The conclusions are limited to synthetic noise on benchmark data.