Kyung Geun Kim

All Publications

• Self-attention with temporal prior: can we learn more from the arrow of time? Frontiers in artificial intelligence Kim, K. G., Lee, B. T. 2024; 7: 1397298

  Abstract

  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.

  View details for DOI 10.3389/frai.2024.1397298

  View details for PubMedID 39165902

  View details for PubMedCentralID PMC11333831

• An Artificial Intelligence Algorithm With 24-h Holter Monitoring for the Identification of Occult Atrial Fibrillation During Sinus Rhythm FRONTIERS IN CARDIOVASCULAR MEDICINE Kim, J., Kim, K., Tae, Y., Chang, M., Park, S., Park, K., On, Y., Kim, J., Lee, Y., Jang, S. 2022; 9: 906780

  Abstract

  Subclinical atrial fibrillation (AF) is one of the pathogeneses of embolic stroke. Detection of occult AF and providing proper anticoagulant treatment is an important way to prevent stroke recurrence. The purpose of this study was to determine whether an artificial intelligence (AI) model can assess occult AF using 24-h Holter monitoring during normal sinus rhythm.This study is a retrospective cohort study that included those who underwent Holter monitoring. The primary outcome was identifying patients with AF analyzed with an AI model using 24-h Holter monitoring without AF documentation. We trained the AI using a Holter monitor, including supraventricular ectopy (SVE) events (setting 1) and excluding SVE events (setting 2). Additionally, we performed comparisons using the SVE burden recorded in Holter annotation data.The area under the receiver operating characteristics curve (AUROC) of setting 1 was 0.85 (0.83-0.87) and that of setting 2 was 0.84 (0.82-0.86). The AUROC of the SVE burden with Holter annotation data was 0.73. According to the diurnal period, the AUROCs for daytime were 0.83 (0.78-0.88) for setting 1 and 0.83 (0.78-0.88) for setting 2, respectively, while those for nighttime were 0.85 (0.82-0.88) for setting 1 and 0.85 (0.80-0.90) for setting 2.We have demonstrated that an AI can identify occult paroxysmal AF using 24-h continuous ambulatory Holter monitoring during sinus rhythm. The performance of our AI model outperformed the use of SVE burden in the Holter exam to identify paroxysmal AF. According to the diurnal period, nighttime recordings showed more favorable performance compared to daytime recordings.

  View details for DOI 10.3389/fcvm.2022.906780

  View details for Web of Science ID 000829444500001

  View details for PubMedID 35872911

  View details for PubMedCentralID PMC9299422

• NeuroGPU: Accelerating multi-compartment, biophysically detailed neuron simulations on GPUs JOURNAL OF NEUROSCIENCE METHODS Ben-Shalom, R., Ladd, A., Artherya, N. S., Cross, C., Kim, K., Sanghevi, H., Korngreen, A., Bouchard, K. E., Bender, K. J. 2022; 366: 109400

  Abstract

  The membrane potential of individual neurons depends on a large number of interacting biophysical processes operating on spatial-temporal scales spanning several orders of magnitude. The multi-scale nature of these processes dictates that accurate prediction of membrane potentials in specific neurons requires the utilization of detailed simulations. Unfortunately, constraining parameters within biologically detailed neuron models can be difficult, leading to poor model fits. This obstacle can be overcome partially by numerical optimization or detailed exploration of parameter space. However, these processes, which currently rely on central processing unit (CPU) computation, often incur orders of magnitude increases in computing time for marginal improvements in model behavior. As a result, model quality is often compromised to accommodate compute resources.Here, we present a simulation environment, NeuroGPU, that takes advantage of the inherent parallelized structure of the graphics processing unit (GPU) to accelerate neuronal simulation.NeuroGPU can simulate most biologically detailed models 10-200 times faster than NEURON simulation running on a single core and 5 times faster than GPU simulators (CoreNEURON). NeuroGPU is designed for model parameter tuning and best performs when the GPU is fully utilized by running multiple (> 100) instances of the same model with different parameters. When using multiple GPUs, NeuroGPU can reach to a speed-up of 800 fold compared to single core simulations, especially when simulating the same model morphology with different parameters. We demonstrate the power of NeuoGPU through large-scale parameter exploration to reveal the response landscape of a neuron. Finally, we accelerate numerical optimization of biophysically detailed neuron models to achieve highly accurate fitting of models to simulation and experimental data.Thus, NeuroGPU is the fastest available platform that enables rapid simulation of multi-compartment, biophysically detailed neuron models on commonly used computing systems accessible by many scientists.

  View details for DOI 10.1016/j.jneumeth.2021.109400

  View details for Web of Science ID 000722687200003

  View details for PubMedID 34728257

  View details for PubMedCentralID PMC9887806

• Scaling and Benchmarking an Evolutionary Algorithm for Constructing Biophysical Neuronal Models FRONTIERS IN NEUROINFORMATICS Ladd, A., Kim, K., Balewski, J., Bouchard, K., Ben-Shalom, R. 2021; 16: 882552

  Abstract

  Single neuron models are fundamental for computational modeling of the brain's neuronal networks, and understanding how ion channel dynamics mediate neural function. A challenge in defining such models is determining biophysically realistic channel distributions. Here, we present an efficient, highly parallel evolutionary algorithm for developing such models, named NeuroGPU-EA. NeuroGPU-EA uses CPUs and GPUs concurrently to simulate and evaluate neuron membrane potentials with respect to multiple stimuli. We demonstrate a logarithmic cost for scaling the stimuli used in the fitting procedure. NeuroGPU-EA outperforms the typically used CPU based evolutionary algorithm by a factor of 10 on a series of scaling benchmarks. We report observed performance bottlenecks and propose mitigation strategies. Finally, we also discuss the potential of this method for efficient simulation and evaluation of electrophysiological waveforms.

  View details for DOI 10.3389/fninf.2022.882552

  View details for Web of Science ID 000891535700001

  View details for PubMedID 35784184

  View details for PubMedCentralID PMC9248031