Search

Your search keyword '"Lee, Hayeon"' showing total 207 results
Start Over Author "Lee, Hayeon" Publication Year Range Last 50 years
207 results on '"Lee, Hayeon"'

1. Diffusion-based Neural Network Weights Generation

Author

Soro, Bedionita, Andreis, Bruno, Lee, Hayeon, Chong, Song, Hutter, Frank, and Hwang, Sung Ju

Subjects
Computer Science - Machine Learning, Computer Science - Artificial Intelligence
Abstract

Transfer learning is a topic of significant interest in recent deep learning research because it enables faster convergence and improved performance on new tasks. While the performance of transfer learning depends on the similarity of the source data to the target data, it is costly to train a model on a large number of datasets. Therefore, pretrained models are generally blindly selected with the hope that they will achieve good performance on the given task. To tackle such suboptimality of the pretrained models, we propose an efficient and adaptive transfer learning scheme through dataset-conditioned pretrained weights sampling. Specifically, we use a latent diffusion model with a variational autoencoder that can reconstruct the neural network weights, to learn the distribution of a set of pretrained weights conditioned on each dataset for transfer learning on unseen datasets. By learning the distribution of a neural network on a variety pretrained models, our approach enables adaptive sampling weights for unseen datasets achieving faster convergence and reaching competitive performance., Comment: 14 pages

Published
2024

2. Psychosocial alterations during the COVID-19 pandemic and the global burden of anxiety and major depressive disorders in adolescents, 1990–2021: challenges in mental health amid socioeconomic disparities

Author

Kim, Soeun, Hwang, Jiyoung, Lee, Jun Hyuk, Park, Jaeyu, Kim, Hyeon Jin, Son, Yejun, Oh, Hans, Smith, Lee, Kang, Jiseung, Fond, Guillaume, Boyer, Laurent, Rahmati, Masoud, Tully, Mark A., Pizzol, Damiano, Udeh, Raphael, Lee, Jinseok, Lee, Hayeon, Lee, Sooji, and Yon, Dong Keon

Published
2024
Full Text
View/download PDF

3. Short- and long-term neuropsychiatric outcomes in long COVID in South Korea and Japan

Author

Kim, Sunyoung, Lee, Hayeon, Lee, Jinseok, Lee, Seung Won, Kwon, Rosie, Kim, Min Seo, Koyanagi, Ai, Smith, Lee, Fond, Guillaume, Boyer, Laurent, Rahmati, Masoud, López Sánchez, Guillermo F., Dragioti, Elena, Cortese, Samuele, Shin, Ju-Young, Choi, Ahhyung, Suh, Hae Sun, Lee, Sunmi, Solmi, Marco, Min, Chanyang, Shin, Jae Il, Yon, Dong Keon, and Fusar-Poli, Paolo

Published
2024
Full Text
View/download PDF

6. Co-training and Co-distillation for Quality Improvement and Compression of Language Models

Author

Lee, Hayeon, Hou, Rui, Kim, Jongpil, Liang, Davis, Zhang, Hongbo, Hwang, Sung Ju, and Min, Alexander

Subjects
Computer Science - Computation and Language, Computer Science - Artificial Intelligence
Abstract

Knowledge Distillation (KD) compresses computationally expensive pre-trained language models (PLMs) by transferring their knowledge to smaller models, allowing their use in resource-constrained or real-time settings. However, most smaller models fail to surpass the performance of the original larger model, resulting in sacrificing performance to improve inference speed. To address this issue, we propose Co-Training and Co-Distillation (CTCD), a novel framework that improves performance and inference speed together by co-training two models while mutually distilling knowledge. The CTCD framework successfully achieves this based on two significant findings: 1) Distilling knowledge from the smaller model to the larger model during co-training improves the performance of the larger model. 2) The enhanced performance of the larger model further boosts the performance of the smaller model. The CTCD framework shows promise as it can be combined with existing techniques like architecture design or data augmentation, replacing one-way KD methods, to achieve further performance improvement. Extensive ablation studies demonstrate the effectiveness of CTCD, and the small model distilled by CTCD outperforms the original larger model by a significant margin of 1.66 on the GLUE benchmark., Comment: Findings of EMNLP 2023

Published
2023

8. A Study on Knowledge Distillation from Weak Teacher for Scaling Up Pre-trained Language Models

Author

Lee, Hayeon, Hou, Rui, Kim, Jongpil, Liang, Davis, Hwang, Sung Ju, and Min, Alexander

Subjects
Computer Science - Computation and Language, Computer Science - Artificial Intelligence
Abstract

Distillation from Weak Teacher (DWT) is a method of transferring knowledge from a smaller, weaker teacher model to a larger student model to improve its performance. Previous studies have shown that DWT can be effective in the vision domain and natural language processing (NLP) pre-training stage. Specifically, DWT shows promise in practical scenarios, such as enhancing new generation or larger models using pre-trained yet older or smaller models and lacking a resource budget. However, the optimal conditions for using DWT have yet to be fully investigated in NLP pre-training. Therefore, this study examines three key factors to optimize DWT, distinct from those used in the vision domain or traditional knowledge distillation. These factors are: (i) the impact of teacher model quality on DWT effectiveness, (ii) guidelines for adjusting the weighting value for DWT loss, and (iii) the impact of parameter remapping as a student model initialization technique for DWT., Comment: Findings of ACL 2023

Published
2023

9. Meta-prediction Model for Distillation-Aware NAS on Unseen Datasets

Author

Lee, Hayeon, An, Sohyun, Kim, Minseon, and Hwang, Sung Ju

Subjects
Computer Science - Machine Learning, Computer Science - Artificial Intelligence
Abstract

Distillation-aware Neural Architecture Search (DaNAS) aims to search for an optimal student architecture that obtains the best performance and/or efficiency when distilling the knowledge from a given teacher model. Previous DaNAS methods have mostly tackled the search for the neural architecture for fixed datasets and the teacher, which are not generalized well on a new task consisting of an unseen dataset and an unseen teacher, thus need to perform a costly search for any new combination of the datasets and the teachers. For standard NAS tasks without KD, meta-learning-based computationally efficient NAS methods have been proposed, which learn the generalized search process over multiple tasks (datasets) and transfer the knowledge obtained over those tasks to a new task. However, since they assume learning from scratch without KD from a teacher, they might not be ideal for DaNAS scenarios. To eliminate the excessive computational cost of DaNAS methods and the sub-optimality of rapid NAS methods, we propose a distillation-aware meta accuracy prediction model, DaSS (Distillation-aware Student Search), which can predict a given architecture's final performances on a dataset when performing KD with a given teacher, without having actually to train it on the target task. The experimental results demonstrate that our proposed meta-prediction model successfully generalizes to multiple unseen datasets for DaNAS tasks, largely outperforming existing meta-NAS methods and rapid NAS baselines. Code is available at https://github.com/CownowAn/DaSS, Comment: ICLR 2023 (Notable-top-25%)

Published
2023

10. DiffusionNAG: Predictor-guided Neural Architecture Generation with Diffusion Models

Author

An, Sohyun, Lee, Hayeon, Jo, Jaehyeong, Lee, Seanie, and Hwang, Sung Ju

Subjects
Computer Science - Machine Learning
Abstract

Existing NAS methods suffer from either an excessive amount of time for repetitive sampling and training of many task-irrelevant architectures. To tackle such limitations of existing NAS methods, we propose a paradigm shift from NAS to a novel conditional Neural Architecture Generation (NAG) framework based on diffusion models, dubbed DiffusionNAG. Specifically, we consider the neural architectures as directed graphs and propose a graph diffusion model for generating them. Moreover, with the guidance of parameterized predictors, DiffusionNAG can flexibly generate task-optimal architectures with the desired properties for diverse tasks, by sampling from a region that is more likely to satisfy the properties. This conditional NAG scheme is significantly more efficient than previous NAS schemes which sample the architectures and filter them using the property predictors. We validate the effectiveness of DiffusionNAG through extensive experiments in two predictor-based NAS scenarios: Transferable NAS and Bayesian Optimization (BO)-based NAS. DiffusionNAG achieves superior performance with speedups of up to 35 times when compared to the baselines on Transferable NAS benchmarks. Furthermore, when integrated into a BO-based algorithm, DiffusionNAG outperforms existing BO-based NAS approaches, particularly in the large MobileNetV3 search space on the ImageNet 1K dataset. Code is available at https://github.com/CownowAn/DiffusionNAG., Comment: Accepted to ICLR 2024

Published
2023

11. A Survey of Supernet Optimization and its Applications: Spatial and Temporal Optimization for Neural Architecture Search

Author

Cha, Stephen, Kim, Taehyeon, Lee, Hayeon, and Yun, Se-Young

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Machine Learning
Abstract

This survey focuses on categorizing and evaluating the methods of supernet optimization in the field of Neural Architecture Search (NAS). Supernet optimization involves training a single, over-parameterized network that encompasses the search space of all possible network architectures. The survey analyses supernet optimization methods based on their approaches to spatial and temporal optimization. Spatial optimization relates to optimizing the architecture and parameters of the supernet and its subnets, while temporal optimization deals with improving the efficiency of selecting architectures from the supernet. The benefits, limitations, and potential applications of these methods in various tasks and settings, including transferability, domain generalization, and Transformer models, are also discussed.

Published
2022

13. Online Hyperparameter Meta-Learning with Hypergradient Distillation

Author

Lee, Hae Beom, Lee, Hayeon, Shin, Jaewoong, Yang, Eunho, Hospedales, Timothy, and Hwang, Sung Ju

Subjects
Computer Science - Machine Learning
Abstract

Many gradient-based meta-learning methods assume a set of parameters that do not participate in inner-optimization, which can be considered as hyperparameters. Although such hyperparameters can be optimized using the existing gradient-based hyperparameter optimization (HO) methods, they suffer from the following issues. Unrolled differentiation methods do not scale well to high-dimensional hyperparameters or horizon length, Implicit Function Theorem (IFT) based methods are restrictive for online optimization, and short horizon approximations suffer from short horizon bias. In this work, we propose a novel HO method that can overcome these limitations, by approximating the second-order term with knowledge distillation. Specifically, we parameterize a single Jacobian-vector product (JVP) for each HO step and minimize the distance from the true second-order term. Our method allows online optimization and also is scalable to the hyperparameter dimension and the horizon length. We demonstrate the effectiveness of our method on two different meta-learning methods and three benchmark datasets.

Published
2021

14. Rapid Neural Architecture Search by Learning to Generate Graphs from Datasets

Author

Lee, Hayeon, Hyung, Eunyoung, and Hwang, Sung Ju

Subjects
Computer Science - Machine Learning, Computer Science - Computer Vision and Pattern Recognition
Abstract

Despite the success of recent Neural Architecture Search (NAS) methods on various tasks which have shown to output networks that largely outperform human-designed networks, conventional NAS methods have mostly tackled the optimization of searching for the network architecture for a single task (dataset), which does not generalize well across multiple tasks (datasets). Moreover, since such task-specific methods search for a neural architecture from scratch for every given task, they incur a large computational cost, which is problematic when the time and monetary budget are limited. In this paper, we propose an efficient NAS framework that is trained once on a database consisting of datasets and pretrained networks and can rapidly search for a neural architecture for a novel dataset. The proposed MetaD2A (Meta Dataset-to-Architecture) model can stochastically generate graphs (architectures) from a given set (dataset) via a cross-modal latent space learned with amortized meta-learning. Moreover, we also propose a meta-performance predictor to estimate and select the best architecture without direct training on target datasets. The experimental results demonstrate that our model meta-learned on subsets of ImageNet-1K and architectures from NAS-Bench 201 search space successfully generalizes to multiple unseen datasets including CIFAR-10 and CIFAR-100, with an average search time of 33 GPU seconds. Even under MobileNetV3 search space, MetaD2A is 5.5K times faster than NSGANetV2, a transferable NAS method, with comparable performance. We believe that the MetaD2A proposes a new research direction for rapid NAS as well as ways to utilize the knowledge from rich databases of datasets and architectures accumulated over the past years. Code is available at https://github.com/HayeonLee/MetaD2A., Comment: ICLR 2021

Published
2021

16. HELP: Hardware-Adaptive Efficient Latency Prediction for NAS via Meta-Learning

Author

Lee, Hayeon, Lee, Sewoong, Chong, Song, and Hwang, Sung Ju

Subjects
Computer Science - Machine Learning
Abstract

For deployment, neural architecture search should be hardware-aware, in order to satisfy the device-specific constraints (e.g., memory usage, latency and energy consumption) and enhance the model efficiency. Existing methods on hardware-aware NAS collect a large number of samples (e.g., accuracy and latency) from a target device, either builds a lookup table or a latency estimator. However, such approach is impractical in real-world scenarios as there exist numerous devices with different hardware specifications, and collecting samples from such a large number of devices will require prohibitive computational and monetary cost. To overcome such limitations, we propose Hardware-adaptive Efficient Latency Predictor (HELP), which formulates the device-specific latency estimation problem as a meta-learning problem, such that we can estimate the latency of a model's performance for a given task on an unseen device with a few samples. To this end, we introduce novel hardware embeddings to embed any devices considering them as black-box functions that output latencies, and meta-learn the hardware-adaptive latency predictor in a device-dependent manner, using the hardware embeddings. We validate the proposed HELP for its latency estimation performance on unseen platforms, on which it achieves high estimation performance with as few as 10 measurement samples, outperforming all relevant baselines. We also validate end-to-end NAS frameworks using HELP against ones without it, and show that it largely reduces the total time cost of the base NAS method, in latency-constrained settings. Code is available at https://github.com/HayeonLee/HELP., Comment: NeurIPS 2021 (Spotlight)

Published
2021

17. Task-Adaptive Neural Network Search with Meta-Contrastive Learning

Author

Jeong, Wonyong, Lee, Hayeon, Park, Gun, Hyung, Eunyoung, Baek, Jinheon, and Hwang, Sung Ju

Subjects
Computer Science - Machine Learning
Abstract

Most conventional Neural Architecture Search (NAS) approaches are limited in that they only generate architectures without searching for the optimal parameters. While some NAS methods handle this issue by utilizing a supernet trained on a large-scale dataset such as ImageNet, they may be suboptimal if the target tasks are highly dissimilar from the dataset the supernet is trained on. To address such limitations, we introduce a novel problem of \emph{Neural Network Search} (NNS), whose goal is to search for the optimal pretrained network for a novel dataset and constraints (e.g. number of parameters), from a model zoo. Then, we propose a novel framework to tackle the problem, namely \emph{Task-Adaptive Neural Network Search} (TANS). Given a model-zoo that consists of network pretrained on diverse datasets, we use a novel amortized meta-learning framework to learn a cross-modal latent space with contrastive loss, to maximize the similarity between a dataset and a high-performing network on it, and minimize the similarity between irrelevant dataset-network pairs. We validate the effectiveness and efficiency of our method on ten real-world datasets, against existing NAS/AutoML baselines. The results show that our method instantly retrieves networks that outperform models obtained with the baselines with significantly fewer training steps to reach the target performance, thus minimizing the total cost of obtaining a task-optimal network. Our code and the model-zoo are available at https://github.com/wyjeong/TANS., Comment: NeurIPS 2021 (Spotlight)

Published
2021

19. Learning to Generalize to Unseen Tasks with Bilevel Optimization

Author

Lee, Hayeon, Na, Donghyun, Lee, Hae Beom, and Hwang, Sung Ju

Subjects
Computer Science - Machine Learning, Statistics - Machine Learning
Abstract

Recent metric-based meta-learning approaches, which learn a metric space that generalizes well over combinatorial number of different classification tasks sampled from a task distribution, have been shown to be effective for few-shot classification tasks of unseen classes. They are often trained with episodic training where they iteratively train a common metric space that reduces distance between the class representatives and instances belonging to each class, over large number of episodes with random classes. However, this training is limited in that while the main target is the generalization to the classification of unseen classes during training, there is no explicit consideration of generalization during meta-training phase. To tackle this issue, we propose a simple yet effective meta-learning framework for metricbased approaches, which we refer to as learning to generalize (L2G), that explicitly constrains the learning on a sampled classification task to reduce the classification error on a randomly sampled unseen classification task with a bilevel optimization scheme. This explicit learning aimed toward generalization allows the model to obtain a metric that separates well between unseen classes. We validate our L2G framework on mini-ImageNet and tiered-ImageNet datasets with two base meta-learning few-shot classification models, Prototypical Networks and Relation Networks. The results show that L2G significantly improves the performance of the two methods over episodic training. Further visualization shows that L2G obtains a metric space that clusters and separates unseen classes well., Comment: 9 pages, 3 figures

Published
2019

20. Learning to Balance: Bayesian Meta-Learning for Imbalanced and Out-of-distribution Tasks

Author

Lee, Hae Beom, Lee, Hayeon, Na, Donghyun, Kim, Saehoon, Park, Minseop, Yang, Eunho, and Hwang, Sung Ju

Subjects
Computer Science - Machine Learning, Statistics - Machine Learning
Abstract

While tasks could come with varying the number of instances and classes in realistic settings, the existing meta-learning approaches for few-shot classification assume that the number of instances per task and class is fixed. Due to such restriction, they learn to equally utilize the meta-knowledge across all the tasks, even when the number of instances per task and class largely varies. Moreover, they do not consider distributional difference in unseen tasks, on which the meta-knowledge may have less usefulness depending on the task relatedness. To overcome these limitations, we propose a novel meta-learning model that adaptively balances the effect of the meta-learning and task-specific learning within each task. Through the learning of the balancing variables, we can decide whether to obtain a solution by relying on the meta-knowledge or task-specific learning. We formulate this objective into a Bayesian inference framework and tackle it using variational inference. We validate our Bayesian Task-Adaptive Meta-Learning (Bayesian TAML) on multiple realistic task- and class-imbalanced datasets, on which it significantly outperforms existing meta-learning approaches. Further ablation study confirms the effectiveness of each balancing component and the Bayesian learning framework.

Published
2019

21. Global association of secondhand smoke exposure locations and smoking behaviour among adolescents in 99 countries

Author

Hong, Seohyun, primary, Son, Yejun, additional, Lee, Hyeri, additional, Kim, Soeun, additional, Kim, Hyeon Jin, additional, Jo, Hyesu, additional, Park, Jaeyu, additional, Lee, Kyeongmin, additional, Lee, Hayeon, additional, Dragioti, Elena, additional, Fond, Guillaume, additional, Boyer, Laurent, additional, López Sánchez, Guillermo F., additional, Tully, Mark A., additional, Rahmati, Masoud, additional, Smith, Lee, additional, Kim, Sunyoung, additional, Woo, Selin, additional, and Yon, Dong Keon, additional

Published
2024
Full Text
View/download PDF

22. Global and regional burden of vaccine‐induced thrombotic thrombocytopenia, 1969–2023: Comprehensive findings with critical analysis of the international pharmacovigilance database

Author

Lee, Sooji, primary, Jo, Hyesu, additional, Woo, Selin, additional, Jeong, Yi Deun, additional, Lee, Hayeon, additional, Lee, Kyeongmin, additional, Lee, Jinseok, additional, Kim, Hyeon Jin, additional, Kang, Jiseung, additional, Jacob, Louis, additional, Smith, Lee, additional, Rahmati, Masoud, additional, López Sánchez, Guillermo F., additional, Dragioti, Elena, additional, Son, Yejun, additional, Kim, Soeun, additional, Yeo, Seung Geun, additional, Park, Jaeyu, additional, and Yon, Dong Keon, additional

Published
2024
Full Text
View/download PDF

25. Global estimates on the reports of vaccine‐associated myocarditis and pericarditis from 1969 to 2023: Findings with critical reanalysis from the WHO pharmacovigilance database

Author

Lee, Sooji, primary, Jo, Hyesu, additional, Lee, Hyeri, additional, Lee, Hayeon, additional, Lee, Jinseok, additional, Kim, Hyeon Jin, additional, Kang, Jiseung, additional, Jacob, Louis, additional, Smith, Lee, additional, Rahmati, Masoud, additional, López Sánchez, Guillermo F., additional, Dragioti, Elena, additional, Jeon, Hongki, additional, Cho, Jin‐Man, additional, Choi, Yujin, additional, Park, Jaeyu, additional, Woo, Selin, additional, and Yon, Dong Keon, additional

Published
2024
Full Text
View/download PDF

26. Acute and post-acute respiratory complications of SARS-CoV-2 infection: population-based cohort study in South Korea and Japan

Author

Choi, Yujin, primary, Kim, Hyeon Jin, additional, Park, Jaeyu, additional, Lee, Myeongcheol, additional, Kim, Sunyoung, additional, Koyanagi, Ai, additional, Smith, Lee, additional, Kim, Min Seo, additional, Rahmati, Masoud, additional, Lee, Hayeon, additional, Kang, Jiseung, additional, and Yon, Dong Keon, additional

Published
2024
Full Text
View/download PDF

27. Global and regional burden of vaccine‐associated facial paralysis, 1967–2023: Findings from the WHO international pharmacovigilance database

Author

Jeong, Yi Deun, primary, Lee, Kyeongmin, additional, Lee, Sooji, additional, Park, Jaeyu, additional, Kim, Hyeon Jin, additional, Lee, Jinseok, additional, Kang, Jiseung, additional, Jacob, Louis, additional, Smith, Lee, additional, Rahmati, Masoud, additional, López Sánchez, Guillermo F., additional, Dragioti, Elena, additional, Son, Yejun, additional, Kim, Soeun, additional, Yeo, Seung Geun, additional, Lee, Hayeon, additional, and Yon, Dong Keon, additional

Published
2024
Full Text
View/download PDF

29. Global burden of vaccine‐associated rheumatic diseases and their related vaccines, 1967–2023: A comprehensive analysis of the international pharmacovigilance database.

Author

Oh, Jiyeon, Jo, Hyesu, Park, Jaeyu, Lee, Hayeon, Kim, Hyeon Jin, Lee, Hyeri, Kang, Jiseung, Hwang, Jiyoung, Woo, Selin, Son, Yejun, Kim, Soeun, Smith, Lee, Rahmati, Masoud, Jacob, Louis, Lee, Jinseok, Lee, Jun Hyuk, López Sánchez, Guillermo F., Dragioti, Elena, Udeh, Raphael, and Veronese, Nicola

Subjects
SJOGREN'S syndrome, RHEUMATISM, DRUG side effects, MEDICAL personnel, SYSTEMIC scleroderma, POLYMYALGIA rheumatica, ANKYLOSING spondylitis
Abstract

Vaccine‐associated rheumatic diseases are rare but one of the most feared adverse drug reactions (ADRs). However, this topic has been investigated less with large‐scale data in the literature. With the rapid progress in the development and approval of vaccines during the pandemic, public concerns regarding their safety have been raised. To assess the global and regional burden, long‐term trends, and potential risk factors of vaccines‐associated six types of rheumatic diseases (ankylosing spondylitis [AS], polymyalgia rheumatica [PMR], rheumatoid arthritis [RA], Sjögren's syndrome, Systemic lupus erythematosus [SLE], Systemic scleroderma), this study conducted disproportionality analysis based on the reports from the World Health Organization International Pharmacovigilance Database documented between 1967 and 2023 (n for total reports = 131 255 418) across 156 countries and territories. We estimated the reporting odds ratio (ROR) and information component (IC) to determine the disproportionality signal for rheumatic diseases. Of 198 046 reports of all‐cause rheumatic diseases, 14 703 reports of vaccine‐associated rheumatic diseases were identified. While the reporting counts have gradually increased over time globally, we observed a dramatic increase in reporting counts after 2020, potentially due to a large portion of reports of COVID‐19 mRNA vaccine‐associated rheumatic diseases. The disproportionality signal for rheumatic diseases was most pronounced in HBV vaccines (ROR, 4.11; IC025, 1.90), followed by COVID‐19 mRNA (ROR, 2.79; IC025, 1.25), anthrax (ROR, 2.52; IC025, 0.76), papillomavirus (ROR, 2.16; IC025, 0.95), encephalitis (ROR, 2.01; IC025, 0.58), typhoid (ROR, 1.91; IC025, 0.44), influenza (ROR, 1.49; IC025, 0.46), and HAV vaccines (ROR, 1.41; IC025, 0.20). From age‐ and sex‐specific perspective, young females and old males are likely to have vaccine‐associated rheumatic disease reports. Furthermore, overall vaccines showed a disproportionality signal for PMR (IC025, 3.13) and Sjögren's syndrome (IC025, 0.70), systemic scleroderma (IC025, 0.64), specifically while the COVID‐19 mRNA vaccines are associated with all six types of diseases. Although multiple vaccines are associated with rheumatic disease reports, healthcare providers should be aware of the potential of autoimmune manifestations following vaccination, particularly the COVID‐19 mRNA and HBV vaccines, and take into account for risk factors associated with these ADRs. Most ADRs exhibited an average time to onset of 11 days, underscoring the significance of monitoring and timely management by clinicians. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

30. Global burden of anticancer drug-induced acute kidney injury and tubulointerstitial nephritis from 1967 to 2023.

Author

Yoon, Soo-Young, Lee, Sooji, Lee, Kyeongmin, Kim, Jin Sug, Hwang, Hyeon Seok, Kronbichler, Andreas, Jacob, Louis, Shin, Ju-Young, Lee, Jin A., Park, Jaeyu, Lee, Hyeri, Lee, Hayeon, Jeong, Kyunghwan, and Yon, Dong Keon

Subjects
ACUTE kidney failure, DRUG side effects, NEPHRITIS, GROUP psychotherapy
Abstract

This study aims to figure out the worldwide prevalence of anticancer therapy-associated acute kidney injury (AKI) and tubulointerstitial nephritis (TIN) and the relative risk of each cancer drug. We conducted an analysis of VigiBase, the World Health Organization pharmacovigilance database, 1967–2023 via disproportionate Bayesian reporting method. We further categorized the anticancer drugs into four groups: cytotoxic therapy, hormone therapy, immunotherapy, and targeted therapy. Reporting odds ratio (ROR) and information component (IC) compares observed and expected values to investigate the associations of each category of anticancer drugs with AKI and TIN. We identified 32,722 and 2056 reports (male, n = 17,829 and 1,293) of anticancer therapy-associated AKI and TIN, respectively, among 4,592,036 reports of all-drug caused AKI and TIN. There has been a significant increase in reports since 2010, primarily due to increased reports of targeted therapy and immunotherapy. Immunotherapy exhibited a significant association with both AKI (ROR: 8.92; IC0.25: 3.06) and TIN (21.74; 4.24), followed by cytotoxic therapy (7.14; 2.68), targeted therapy (5.83; 2.40), and hormone therapy (2.59; 1.24) for AKI, and by cytotoxic therapy (2.60; 1.21) and targeted therapy (1.54; 0.61) for TIN. AKI and TIN were more prevalent among individuals under 45 years of age, with a female preponderance for AKI and males for TIN. These events were reported in close temporal relationship after initiation of the respective drug (16.53 days for AKI and 27.97 days for TIN), and exhibited a high fatality rate, with 23.6% for AKI and 16.3% for TIN. These findings underscore that kidney-related adverse drug reactions are of prognostic significance and strategies to mitigate such side effects are required to optimize anticancer therapy. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

31. GolpHCat (TMEM87A), a unique voltage-dependent cation channel in Golgi apparatus, contributes to Golgi-pH maintenance and hippocampus-dependent memory.

Author

Kang, Hyunji, Han, Ah-reum, Zhang, Aihua, Jeong, Heejin, Koh, Wuhyun, Lee, Jung Moo, Lee, Hayeon, Jo, Hee Young, Maria-Solano, Miguel A., Bhalla, Mridula, Kwon, Jea, Roh, Woo Suk, Yang, Jimin, An, Hyun Joo, Choi, Sun, Kim, Ho Min, and Lee, C. Justin

Subjects
GOLGI apparatus, HIPPOCAMPUS (Brain), ION channels, MEMBRANE proteins, SPATIAL memory, ALZHEIMER'S disease, COGNITIVE structures
Abstract

Impaired ion channels regulating Golgi pH lead to structural alterations in the Golgi apparatus, such as fragmentation, which is found, along with cognitive impairment, in Alzheimer's disease. However, the causal relationship between altered Golgi structure and cognitive impairment remains elusive due to the lack of understanding of ion channels in the Golgi apparatus of brain cells. Here, we identify that a transmembrane protein TMEM87A, renamed Golgi-pH-regulating cation channel (GolpHCat), expressed in astrocytes and neurons that contributes to hippocampus-dependent memory. We find that GolpHCat displays unique voltage-dependent currents, which is potently inhibited by gluconate. Additionally, we gain structural insights into the ion conduction through GolpHCat at the molecular level by determining three high-resolution cryogenic-electron microscopy structures of human GolpHCat. GolpHCat-knockout mice show fragmented Golgi morphology and altered protein glycosylation and functions in the hippocampus, leading to impaired spatial memory. These findings suggest a molecular target for Golgi-related diseases and cognitive impairment. Impaired ion channels regulating Golgi pH may cause altered Golgi morphology. Here authors identify TMEM87A as a pH-regulation cation channel expressed in astrocytes and neurons and determine the cryo-EM structure to understand the ion permeation pathway. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

32. Global burden of vaccine‐associated hepatobiliary and gastrointestinal adverse drug reactions, 1967–2023: A comprehensive analysis of the international pharmacovigilance database.

Author

Lee, Sooji, Lee, Kyeongmin, Park, Jaeyu, Jeong, Yi Deun, Jo, Hyesu, Kim, Soeun, Woo, Selin, Son, Yejun, Kim, Hyeon Jin, Lee, Kwanjoo, Ha, Yeonjung, Oh, Na‐eun, Lee, Jinseok, Rhee, Sang Youl, Smith, Lee, Kang, Jiseung, Rahmati, Masoud, Lee, Hayeon, and Yon, Dong Keon

Subjects
DRUG side effects, GASTROINTESTINAL agents, MEDICAL personnel, DATABASES, HEPATITIS A vaccines, HEPATITIS B, CHOLANGITIS
Abstract

Although previous studies have focused on hepatobiliary and gastrointestinal adverse drug reactions (ADRs) associated with COVID‐19 vaccines, literature on such ADRs with other vaccines is limited, particularly on a global scale. Therefore, we aimed to investigate the global burden of vaccine‐associated hepatobiliary and gastrointestinal ADRs and identify the vaccines implicated in these occurrences. This study utilized data from the World Health Organization (WHO) international pharmacovigilance database to extract reports of vaccine‐associated hepatobiliary and gastrointestinal ADRs from 1967 to 2023 (total reports = 131 255 418). Through global reporting counts, reported odds ratios (ROR) with 95% confidence interval (CI), and information components (IC) with IC0.25, the study examined the association between 16 vaccines and the incidence of hepatobiliary and gastrointestinal ADRs across 156 countries. Of the 6 842 303 reports in the vaccine‐associated ADRs, 10 786 reports of liver injury, 927 870 reports of gastrointestinal symptoms, 2978 reports of pancreas and bile duct injury, and 96 reports of intra‐abdominal hemorrhage between 1967 and 2023 were identified. Most hepatobiliary and gastrointestinal ADRs surged after 2020, with the majority of reports attributed to COVID‐19 messenger RNA (mRNA) vaccines. Hepatitis A vaccines exhibited the highest association with liver injury (ROR [95% CI]: 10.30 [9.65–10.99]; IC [IC0.25]: 3.33 [3.22]), followed by hepatitis B, typhoid, and rotavirus. Specifically, ischemic hepatitis had a significant association with both Ad5‐vectored and mRNA COVID‐19 vaccines. Gastrointestinal symptoms were associated with all vaccines except for tuberculosis vaccines, particularly with rotavirus (11.62 [11.45–11.80]; 3.05 [3.03]) and typhoid (11.02 [10.66–11.39]; 3.00 [2.96]). Pancreas and bile duct injury were associated with COVID‐19 mRNA (1.99 [1.89–2.09]; 0.90 [0.83]), MMR (measles, mumps, and rubella), and papillomavirus vaccines. For intra‐abdominal hemorrhage, inactivated whole‐virus COVID‐19 vaccines (3.93 [1.86–8.27]; 1.71 [0.41]) had the highest association, followed by COVID‐19 mRNA (1.81 [1.42–2.29]; 0.77 [0.39]). Most of these ADRs had a short time to onset, within 1 day, and low mortality rate. Through a global scale database, the majority of ADRs occurred within 1 day, emphasizing the importance of healthcare workers' vigilant monitoring and timely management. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

33. Long-Term Autoimmune Inflammatory Rheumatic Outcomes of COVID-19

Author

Kim, Min Seo, primary, Lee, Hayeon, additional, Lee, Seung Won, additional, Kwon, Rosie, additional, Rhee, Sang Youl, additional, Lee, Jin A., additional, Koyanagi, Ai, additional, Smith, Lee, additional, Fond, Guillaume, additional, Boyer, Laurent, additional, Lee, Jinseok, additional, Rahmati, Masoud, additional, Shin, Ju-Young, additional, Min, Chanyang, additional, Shin, Jae Il, additional, and Yon, Dong Keon, additional

Published
2024
Full Text
View/download PDF

34. Machine Learning–Based Prediction of Suicidality in Adolescents With Allergic Rhinitis: Derivation and Validation in 2 Independent Nationwide Cohorts

Author

Lee, Hojae, primary, Cho, Joong Ki, additional, Park, Jaeyu, additional, Lee, Hyeri, additional, Fond, Guillaume, additional, Boyer, Laurent, additional, Kim, Hyeon Jin, additional, Park, Seoyoung, additional, Cho, Wonyoung, additional, Lee, Hayeon, additional, Lee, Jinseok, additional, and Yon, Dong Keon, additional

Published
2024
Full Text
View/download PDF

36. Risks of cutaneous immune‐related adverse events in long COVID: Multinational cohort studies in South Korea, Japan, and the UK.

Author

Kim, Hyejun, Kyung, Seoyeon, Park, Jaeyu, Lee, Hojae, Lee, Myeongcheol, Smith, Lee, Rahmati, Masoud, Shin, Ju‐Young, Kang, Jiseung, Jacob, Louis, Papadopoulos, Nikolaos G., Rhee, Sang Youl, Lee, Jinseok, Kim, Hyeon Jin, Lee, Hayeon, and Yon, Dong Keon

Subjects
POST-acute COVID-19 syndrome, DRUG side effects, PROPORTIONAL hazards models
Abstract

Previous research has not investigated the persistent cutaneous immune‐related adverse events (cirAEs) related to long COVID to investigate the long‐term sequelae. This multinational study, using a propensity‐matched overlap weighting method, utilizes large national claims‐based cohorts, using ICD‐10 code diagnosis, focusing on patients aged ≥20 years from three countries: South Korean, Japanese, and the British cohorts. To estimate the risk of cirAEs in long COVID, the persistence or emergence of cirAEs occurring 4 weeks after the initial SARS‐CoV‐2 infection, we employed a Cox proportional hazard regression model. The Korean cohort (n = 5,937,373; mean age 49.2 years [SD: 13.2]), the Japanese cohort (n = 4,307,587; 42.5 years [13.6]), and the UK cohort (n = 395,435; 71.0 years [8.07]) were presented. An increased risk of cirAEs in long COVID was observed (HR, 1.10; 95% CI, 1.06–1.14) in Korean cohort, while a similar association was observed in Japanese and UK cohorts. The long‐term risk of cirAEs in long COVID was higher in more severe COVID‐19 cases (1.31; 1.22–1.39). Unlike the increased risk of cirAEs in long COVID, COVID‐19 vaccination attenuated the risk, especially with two or more doses (1.03; 0.95–1.11) or heterologous regimens (0.98; 0.76–1.27). The time attenuation effect indicated a sustained risk for up to 6 months postinfection (<3 months: 1.13 [1.07–1.19]; 3‐6 months: 1.14 [1.06‐1.22]). SARS‐CoV‐2 infection is associated with an increased risk of cirAEs in the aspect of long COVID. Vaccination might reduce this risk, highlighting the need for preventive strategies in long COVID management. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

37. Machine learning-based model to predict delirium in patients with advanced cancer treated with palliative care: a multicenter, patient-based registry cohort.

Author

Kim, Yu Jung, Lee, Hayeon, Woo, Ho Geol, Lee, Si Won, Hong, Moonki, Jung, Eun Hee, Yoo, Shin Hye, Lee, Jinseok, Yon, Dong Keon, and Kang, Beodeul

Abstract

This study aimed to present a new approach to predict to delirium admitted to the acute palliative care unit. To achieve this, this study employed machine learning model to predict delirium in patients in palliative care and identified the significant features that influenced the model. A multicenter, patient-based registry cohort study in South Korea between January 1, 2019, and December 31, 2020. Delirium was identified by reviewing the medical records based on the criteria of the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition. The study dataset included 165 patients with delirium among 2314 patients with advanced cancer admitted to the acute palliative care unit. Seven machine learning models, including extreme gradient boosting, adaptive boosting, gradient boosting, light gradient boosting, logistic regression, support vector machine, and random forest, were evaluated to predict delirium in patients with advanced cancer admitted to the acute palliative care unit. An ensemble approach was adopted to determine the optimal model. For k-fold cross-validation, the combination of extreme gradient boosting and random forest provided the best performance, achieving the following accuracy metrics: 68.83% sensitivity, 70.85% specificity, 69.84% balanced accuracy, and 74.55% area under the receiver operating characteristic curve. The performance of the isolated testing dataset was also validated, and the machine learning model was successfully deployed on a public website () to provide public access to delirium prediction results in patients with advanced cancer. Furthermore, using feature importance analysis, sex was determined to be the top contributor in predicting delirium, followed by a history of delirium, chemotherapy, smoking status, alcohol consumption, and living with family. Based on a large-scale, multicenter, patient-based registry cohort, a machine learning prediction model for delirium in patients with advanced cancer was developed in South Korea. We believe that this model will assist healthcare providers in treating patients with delirium and advanced cancer. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

38. Global burden of vaccine-associated alopecia, 1979–2023: a comprehensive analysis of the international pharmacovigilance database.

Author

Kyung, Seoyeon, Woo, Selin, Kim, Minji, Lee, Hayeon, Kang, Jiseung, Rahmati, Masoud, and Yon, Dong Keon

Subjects
ALOPECIA areata, AUTOIMMUNE diseases, DATABASES, BALDNESS, HEPATITIS B vaccines, HUMAN papillomavirus vaccines, HEPATITIS A vaccines
Abstract

A comprehensive analysis of the international pharmacovigilance database, VigiBase, was conducted to examine the global burden of vaccine-associated alopecia. The study analyzed data from 1979 to 2023 and included reports from over 156 countries. The results showed that vaccine-associated alopecia was highly linked to COVID-19 mRNA vaccines, papillomavirus vaccines, and Ad5-vectored COVID-19 vaccines. The study also found that reports of vaccine-associated alopecia were more prevalent among males and showed a U-shaped association with age. However, the study acknowledged limitations, such as underrepresentation of vaccine-associated alopecia from low-income countries. The findings highlight the need for strengthening vaccine safety measures and enhancing adverse-event monitoring. [Extracted from the article]

Published
2024
Full Text
View/download PDF

39. Risks of alopecia areata in long COVID: Binational population‐based cohort studies from South Korea and Japan.

Author

Kyung, Seoyeon, Son, Yejun, Kim, Minji, Kang, Jiseung, Smith, Lee, Lee, Hayeon, and Yon, Dong Keon

Subjects
POST-acute COVID-19 syndrome, ALOPECIA areata, NOSOLOGY, COVID-19, PROPENSITY score matching
Abstract

Previous studies have proposed alopecia areata (AA) as a potential outcome of COVID‐19 infection among autoimmune diseases, yet the findings might be inconclusive and difficult to generalize due to limited sample sizes and evidence levels. Thus, we aimed to investigate in detail the long‐term risk of AA following SARS‐CoV‐2 infection based on large, binational, general population‐based cohort studies. Our study investigated the long‐term AA risk after SARS‐CoV‐2 infection by analyzing bi‐national, claim‐based cohorts in South Korea and Japan: a Korean nationwide cohort (K‐COV‐N cohort; discovery cohort; total n = 10 027 506) and a Japanese claims‐based cohort (JMDC cohort; validation cohort; total n = 12 218 680). AA was identified based on the international classification of diseases 10th revision code (L63) requiring at least three claims within 1 year. After exposure‐driven propensity score matching, SARS‐CoV‐2 infection was associated with an increased risk of incident AA (aHR, 1.66; 95% CI, 1.38–1.99). This increased risk was observed and persisted for up to 6 months. A similar pattern was observed in the validation cohort. As modifiable factors, severe COVID‐19 increased the risk of AA, whereas receiving two or more doses of the COVID‐19 vaccine before infection decreased the risk of AA. Through a bi‐national cohort study in South Korea and Japan, SARS‐CoV‐2 infection was associated with an elevated risk for incident AA in the aspect of long COVID. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

40. Stabilization of Oxygen‐Dependent Fe3+/4+ Redox in Li‐Excess DRX Cathode Exhibiting Anionic Redox via Transition Metal Combination.

Author

Lee, Hayeon, Kim, Minji, Park, Hyunyoung, Yoo, Yiseul, Na, Sangmun, Lim, Hee‐Dae, Kim, Jongsoon, and Yoon, Won‐Sub

Subjects
*TRANSITION metals, *OXIDATION-reduction reaction, *CATHODES, *CHARGE transfer, *REDUCTION potential
Abstract

Developing sustainable Li‐ion batteries requires high‐energy cathodes based on low‐cost, earth‐abundant elements, moving away from low‐reserve nickel and cobalt. Fe‐based oxide cathodes with Fe3+/4+ and O2−/n− redox couples offer potential but face low initial Coulombic efficiency and significant voltage hysteresis. This study investigates Li‐excess Fe‐based disordered rock‐salt (DRX) oxyfluorides (Li2Fe0.5M0.5O2F; M = Fe, Ti, Mn) using combined electrochemical/spectroscopic characterization and first‐principles calculation. Oxygen‐dependent Fe3+/4+ redox, related to Fe 3d–O 2p hybrid state, can be stabilized when combined with Mn3+/4+ redox in DRX structure owing to the unusual decrease in its redox potential. The moderately high charge transfer gap stabilizes Fe4+ against ligand‐to‐metal charge transfer (LMCT) on charge, reduces the amount of oxygen oxidation, thereby increasing Coulombic efficiency. On discharge, it allows metal‐to‐ligand charge transfer (MLCT) without substantial overpotential, reducing hysteresis in oxygen redox. The resulting composition exhibits high capacity (309 mAh g−1) and energy density (998 Wh kg−1), providing insights for next‐generation Ni‐ and Co‐free cathode materials. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

41. Global burden of vaccine‐associated multiple sclerosis, 1967–2022: A comprehensive analysis of the international pharmacovigilance database.

Author

Woo, Ho Geol, Kim, Hyeon Jin, Park, Jaeyu, Lee, Jinseok, Lee, Hayeon, Kim, Min Seo, Koyanagi, Ai, Smith, Lee, Rahmati, Masoud, Yeo, Seung Geun, and Yon, Dong Keon

Subjects
HEPATITIS B, HEPATITIS B vaccines, MULTIPLE sclerosis, DATABASES, VACCINATION complications, HUMAN papillomavirus vaccines
Abstract

Vaccine‐associated multiple sclerosis (MS) is rare, with insufficient evidence from case reports. Given the scarcity of large‐scale data investigating the association between vaccine administration and adverse events, we investigated the global burden of vaccine‐associated MS and potential related vaccines from 1967 to 2022. Reports on vaccine‐associated MS between 1967 and 2022 were obtained from the World Health Organization International Pharmacovigilance Database (total number of reports = 120 715 116). We evaluated global reports, reporting odds ratio (ROR), and information components (IC) to investigate associations between 19 vaccines and vaccine‐associated MS across 156 countries and territories. We identified 8288 reports of vaccine‐associated MS among 132 980 cases of all‐cause MS. The cumulative number of reports on vaccine‐associated MS gradually increased over time, with a substantial increase after 2020, owing to COVID‐19 mRNA vaccine‐associated MS. Vaccine‐associated MS develops more frequently in males and adolescents. Nine vaccines were significantly associated with higher MS reporting, and the highest disproportional associations were observed for hepatitis B vaccines (ROR 19.82; IC025 4.18), followed by encephalitis (ROR 7.42; IC025 2.59), hepatitis A (ROR 4.46; IC025 1.95), and papillomavirus vaccines (ROR 4.45; IC025 2.01). Additionally, MS showed a significantly disproportionate signal for COVID‐19 mRNA vaccines (ROR 1.55; IC025 0.52). Fatal clinical outcomes were reported in only 0.3% (21/8288) of all cases of vaccine‐associated MS. Although various vaccines are potentially associated with increased risk of MS, we should be cautious about the increased risk of MS following vaccination, particularly hepatitis B and COVID‐19 mRNA vaccines, and should consider the risk factors associated with vaccine‐associated MS. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

43. Global burden of vaccine‐associated anaphylaxis and their related vaccines, 1967–2023: A comprehensive analysis of the international pharmacovigilance database

Author

Lee, Kyeonghun, primary, Lee, Hayeon, additional, Kwon, Rosie, additional, Shin, Youn Ho, additional, Yeo, Seung Geun, additional, Lee, Young Joo, additional, Kim, Min Seo, additional, Choi, Yong Sung, additional, Papadopoulos, Nikolaos G., additional, Rahmati, Masoud, additional, Jung, Junyang, additional, Lee, Jinseok, additional, and Yon, Dong Keon, additional

Published
2023
Full Text
View/download PDF

44. Hospital Pharmacists’ Patient-Centered Communication, Trust, and Satisfaction: Patients’ Perceptions and Opinions

Author

Jin, Hye Kyung, primary, Suh, Sung Yun, additional, Kim, Sung Hwan, additional, Kim, A Jeong, additional, Kim, Hyun Jeong, additional, Choi, Nigh, additional, Choi, Jeong Yun, additional, Lee, Hayeon, additional, Hong, Jiseung, additional, Lee, Oksang, additional, Oh, Hyeri, additional, Jeon, Ah Young, additional, Kim, Gahyun, additional, Do, Jihyeon, additional, Lee, Yumi, additional, and Cho, Yoon Sook, additional

Published
2023
Full Text
View/download PDF

48. Comorbid health conditions in people with attention-deficit/hyperactivity disorders: An umbrella review of systematic reviews and meta-analyses

Author

Kang, Jiseung, Lee, Hyeri, Kim, Soeun, Kim, Hyeon Jin, Lee, Hayeon, Kwon, Rosie, Son, Yejun, Kim, Sunyoung, Woo, Ho Geol, Kim, Min Seo, Koyanagi, Ai, Smith, Lee, Fond, Guillaume, Boyer, Laurent, Rahmati, Masoud, López Sánchez, Guillermo F., Dragioti, Elena, Solmi, Marco, Shin, Jae Il, Kim, Tae, Yon, Dong Keon, and Cortese, Samuele

Published
2024
Full Text
View/download PDF

50. National trends in alcohol consumption, smoking, suicide attempts, and COVID-19 pandemic-related factors among South Korean adolescents, 2012–2022

Author

Cheong, Chanyoung, Park, Jaeyu, Park, Junjae, Jo, Hyesu, Lee, Kyeongmin, Lee, Jun Hyuk, Kim, Hyeon Jin, Choi, Yujin, Kim, Hyejun, Nehs, Christa J., Fond, Guillaume, Boyer, Laurent, Rahmati, Masoud, Koyanagi, Ai, Smith, Lee, López Sánchez, Guillermo Felipe, Dragioti, Elena, Woo, Selin, Lee, Hayeon, Kang, Jiseung, Kim, Tae, and Yon, Dong Keon

Published
2024
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results