KKKyung Geun Kim

Kyung Geun Kim

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Kyung Geun Kim has a total of 23 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Accurately fitting biophysical neuron models to experimental voltage data enabled by meta-learning

    by Roy Ben-Shalom, Kyung Geun Kim, Alexander Ladd, Prarthan Ghosh, Timothy Fenton, Henry Kyoung, Ghazaleh Yazdani, Kaustubh Chakravarthula, Gunther Weber, Caleb Geniesse, et al.

    The neuron is a fundamental unit of computation in the brain. Neuronal firing properties are governed by ion channels distributed across complex morphologies. Determining ionic conductances from experimental observations of intracellular membrane potentials would advance basic understanding and clinical translation, but is currently considered intractable. To enable accurate fitting of biophysical neuron models to experimental somatic voltage recordings, we developed a meta-learning algorithm (‘CoMParE’). Using CoMParE to fit standard biophysically detailed models, we demonstrate state-of-the-art reproduction of experimental data and quantify parameter estimate precision and accuracy. We additionally fit models with enhanced electrophysiological detail and demonstrate further improvements in experimental data reproduction. Finally, we show that these improvements result from convexifying the objective function loss surface through meta-learning. These results demonstrate that highly detailed biophysical models can accurately reproduce experimental data, enabling determination of ionic conductances underlying neuronal function and neurological disorders, and indicating that this inverse problem is tractable.

  • Open Access

    Deep Learning-Based Risk Assessment and Prediction of Cardiac Outcomes Using Single-Lead 24-Hour Holter-ECG in Patients with Heart Failure or Myocardial Infarction

    by Ju Youn Kim, Kyung Geun Kim, Sunghoon Joo, Mineok Chang, Juwon Kim, Kyoung-Min Park, Young Keun On, June Soo Kim, Young Soo Lee, Seung-Jung Park

    Background: Deep learning (DL) models using Holter-ECG may enhance risk stratification after heart failure (HF) or myocardial infarction (MI). Objective: To evaluate the prognostic performance of a Holter-based DL model for predicting major adverse cardiac events (MACE), compared with conventional noninvasive markers. Methods: In the K-REDEFINE study, 1108 patients with acute MI or HF underwent 24 h Holter monitoring. A DL model was trained using raw Holter-ECG data and tested for predicting a composite of cardiac death and ventricular arrhythmias. Its performance was compared with heart rate turbulence (HRT), T-wave alternans (TWA), and ejection fraction (EF). Results: During follow-up, 56 adjudicated cardiac deaths (1.18%/yr) and 21 ventricular arrhythmias (0.44%/yr) occurred. The DL model showed an area under the receiver operating characteristic curve (AUROC) of 0.74 (95% CI, 0.70–0.77) for the composite outcome, improving to 0.77 (0.74–0.81) when combined with EF. In comparison, HRT and TWA showed lower AUROCs of 0.62 and 0.55, respectively. For cardiac death alone, the AUROC reached 0.79, further improving to 0.82 with EF. Model-derived risk stratification revealed a seven-fold increase in cardiac death risk in the high-risk group compared to the low-risk group (HR 7.47, 95% CI 2.24–24.96, p < 0.001). This stratification remained particularly effective in patients with EF > 40%. Conclusions: A DL algorithm trained on single-lead Holter-ECG data effectively predicted cardiac death and ventricular arrhythmia. Its performance surpassed conventional markers and was further enhanced when integrated with EF, supporting its potential for noninvasive, scalable risk stratification.


  • Open Access

    Graph structure based data augmentation method

    by Kyung Geun Kim, Byeong Tak Lee

    In this paper, we propose a novel graph-based data augmentation method that can generally be applied to medical waveform data with graph structures. In the process of recording medical waveform data, such as electrocardiogram (ECG), angular perturbations between the measurement leads exist due to imperfections in lead positions. The data samples with large angular perturbations often cause inaccuracy in algorithmic prediction tasks. We design a graph-based data augmentation technique that exploits the inherent graph structures within the medical waveform data to improve the F1 score by 1.44% over various tasks, models, and datasets. In addition, we show that Graph Augmentation improves model robustness by testing against adversarial attacks. Since Graph Augmentation is methodologically orthogonal to existing data augmentation techniques, they can be used in conjunction to further improve the final performance, resulting in a 2.47% gain of the F1 score. We believe that our Graph Augmentation method opens up new possibilities to explore in data augmentation.


  • Open Access

    Self-attention with temporal prior: can we learn more from the arrow of time?

    by Kyung Geun Kim, Byeong Tak Lee

    Many diverse phenomena in nature often inherently encode both short- and long-term temporal dependencies, which especially result from the direction of the flow of time. In this respect, we discovered experimental evidence suggesting that interrelations of these events are higher for closer time stamps. However, to be able for attention-based models to learn these regularities in short-term dependencies, it requires large amounts of data, which are often infeasible. This is because, while they are good at learning piece-wise temporal dependencies, attention-based models lack structures that encode biases in time series. As a resolution, we propose a simple and efficient method that enables attention layers to better encode the short-term temporal bias of these data sets by applying learnable, adaptive kernels directly to the attention matrices. We chose various prediction tasks for the experiments using Electronic Health Records (EHR) data sets since they are great examples with underlying long- and short-term temporal dependencies. Our experiments show exceptional classification results compared to best-performing models on most tasks and data sets.