1,720 Hits in 2.3 sec

Progressive Domain Adaptation for Object Detection [article]

Han-Kai Hsu, Chun-Han Yao, Yi-Hsuan Tsai, Wei-Chih Hung, Hung-Yu Tseng, Maneesh Singh, Ming-Hsuan Yang
2019 arXiv   pre-print
Instead of performing alignment in the feature/pixel space, Tsai et al. [29, 30] adopt adversarial learning in the structured output space for solving domain adaptation on semantic segmentation.  ... 
arXiv:1910.11319v1 fatcat:qd4yzdqmqfanhnv74pd4ftmaa4

Chang Gung Research Database: A multi-institutional database consisting of original medical records

Ming-Shao Tsai, Meng-Hung Lin, Chuan-Pin Lee, Yao-Hsu Yang, Wen-Cheng Chen, Geng-He Chang, Yao-Te Tsai, Pau-Chung Chen, Ying-Huang Tsai
2017 Biomedical Journal  
doi:10.1016/ pmid:29179881 pmcid:PMC6138604 fatcat:254627yaxndqdgsjrvzdwjrf5e

Generative-Discriminative Variational Model for Visual Recognition [article]

Chih-Kuan Yeh and Yao-Hung Hubert Tsai and Yu-Chiang Frank Wang
2017 arXiv   pre-print
The paradigm shift from shallow classifiers with hand-crafted features to end-to-end trainable deep learning models has shown significant improvements on supervised learning tasks. Despite the promising power of deep neural networks (DNN), how to alleviate overfitting during training has been a research topic of interest. In this paper, we present a Generative-Discriminative Variational Model (GDVM) for visual classification, in which we introduce a latent variable inferred from inputs for
more » ... rom inputs for exhibiting generative abilities towards prediction. In other words, our GDVM casts the supervised learning task as a generative learning process, with data discrimination to be jointly exploited for improved classification. In our experiments, we consider the tasks of multi-class classification, multi-label classification, and zero-shot learning. We show that our GDVM performs favorably against the baselines or recent generative DNN models.
arXiv:1706.02295v1 fatcat:gnu5xpmprjf4rfberm4mvmsuoy

Learning Robust Visual-Semantic Embeddings [article]

Yao-Hung Hubert Tsai and Liang-Kang Huang and Ruslan Salakhutdinov
2017 arXiv   pre-print
Many of the existing methods for learning joint embedding of images and text use only supervised information from paired images and its textual attributes. Taking advantage of the recent success of unsupervised learning in deep neural networks, we propose an end-to-end learning framework that is able to extract more robust multi-modal representations across domains. The proposed method combines representation learning models (i.e., auto-encoders) together with cross-domain learning criteria
more » ... arning criteria (i.e., Maximum Mean Discrepancy loss) to learn joint embeddings for semantic and visual features. A novel technique of unsupervised-data adaptation inference is introduced to construct more comprehensive embeddings for both labeled and unlabeled data. We evaluate our method on Animals with Attributes and Caltech-UCSD Birds 200-2011 dataset with a wide range of applications, including zero and few-shot image recognition and retrieval, from inductive to transductive settings. Empirically, we show that our framework improves over the current state of the art on many of the considered tasks.
arXiv:1703.05908v2 fatcat:bmvr3bbvavepbg7k7gwgks7gne

SeqsLab: an integrated platform for cohort-based annotation and interpretation of genetic variants on Spark [article]

Ming-Tai Chang, Yi-An Tung, Jen-Ming Chung, Hung-Fei Yao, Yun-Lung Li, Yin-Hung Lin, Yao-Ting Wang, Chien-Yu Chen, Chung-Tsai Su
2017 bioRxiv   pre-print
SeqsLab uses the Django framework to construct the web server and integrates jBrowse (Buels, Yao et al. 2016) as the genome viewer to show the locations of a selected variant and its relationships to  ... 
doi:10.1101/239962 fatcat:zyfbffioo5dibj7abepllzj3mi

Predictors for gram-negative monomicrobial necrotizing fasciitis in southern Taiwan

Tsung-Yu Huang, Kuo-Ti Peng, Cheng-Ting Hsiao, Wen-Chih Fann, Yao-Hung Tsai, Yen-Yao Li, Chien-Hui Hung, Fang-Yi Chuang, Wei-Hsiu Hsu
2020 BMC Infectious Diseases  
Necrotizing fasciitis (NF) is a rare and life-threatening necrotizing skin and soft-tissue infection. Infectious pathogens of NF must be detected early and treated rapidly to prevent loss of limb or a fatal outcome. This study aimed to detect more reliable predictors between gram-negative and gram-positive monomicrobial NF of limbs. A total of 100 patients with limb monomicrobial NF were diagnosed prospectively from April 2015 to July 2018. These monomicrobial NF pathogens can be divided into
more » ... n be divided into gram-negative and gram-positive groups according to the result of Gram staining and final bacterial reports. Data such as demographics, seawater or seafood contact history, infectious location, comorbidities, presenting signs and symptoms, and laboratory findings were recorded and compared. A total of 55 patients were infected with gram-negative organisms and 45 patients with gram-positive organisms. Among the 55 cases of monomicrobial gram-negative NF, 48 (87.3%) were caused mainly by Vibrio spp. (38, 69.1%) and Aeromonas spp. (10, 18.2%). A higher incidence of chronic kidney disease, cerebrovascular accident, tachypnea, and septic shock; a higher rate of band forms of leukocytes of more than 3%, serum lactate of more than 20 mg/dL, and C-reactive protein level of less than 150 mg/dL; prolonged prothrombin time; and a lower fibrinogen level were observed in patients with gram-negative infection. In a multivariate analysis, a higher incidence of seawater or seafood contact history (odds ratio [OR]: 66.301; 95% confidence interval [CI]: 7.467-588.702), a higher rate of hyperlactatemia (OR: 7.904; 95% CI: 1.231-50.744), and a low fibrinogen level (OR: 1.013; 95% CI: 1.004-1.023) indicated gram-negative infection. In southern Taiwan, NF of limbs mainly affected the lower limbs, exhibited monomicrobial infection, and was predominated by gram-negative bacteria. Gram-negative monomicrobial NF of limbs often occurred in individuals with the more seawater or seafood contact history, hyperlactatemia, and low fibrinogen levels.
doi:10.1186/s12879-020-4796-3 pmid:31959118 fatcat:47w3qmgn4rgancyuxd26bkxl5a

Strong and Simple Baselines for Multimodal Utterance Embeddings [article]

Paul Pu Liang, Yao Chong Lim, Yao-Hung Hubert Tsai, Ruslan Salakhutdinov, Louis-Philippe Morency
2020 arXiv   pre-print
., 2017) , enforcing disentanglement on multimodal representations (Tsai et al., 2018) , and using attention to weight modalities (Gulrajani et al., 2017) led to better performing multimodal representations  ... 
arXiv:1906.02125v2 fatcat:olp2n6ewqvg2zcuy3wo4twas6i

"Dependency Bottleneck" in Auto-encoding Architectures: an Empirical Study [article]

Denny Wu, Yixiu Zhao, Yao-Hung Hubert Tsai, Makoto Yamada, Ruslan Salakhutdinov
2018 arXiv   pre-print
Recent works investigated the generalization properties in deep neural networks (DNNs) by studying the Information Bottleneck in DNNs. However, the mea- surement of the mutual information (MI) is often inaccurate due to the density estimation. To address this issue, we propose to measure the dependency instead of MI between layers in DNNs. Specifically, we propose to use Hilbert-Schmidt Independence Criterion (HSIC) as the dependency measure, which can measure the dependence of two random
more » ... of two random variables without estimating probability densities. Moreover, HSIC is a special case of the Squared-loss Mutual Information (SMI). In the experiment, we empirically evaluate the generalization property using HSIC in both the reconstruction and prediction auto-encoding (AE) architectures.
arXiv:1802.05408v1 fatcat:sltamp7gdba3ngmvu2n5bvpxzm

Interactive Task Assignment Model for Internet of Things Devices

Tse-Chuan Hsu, Yao-Hong Tsai, Dong-Meau Chang, Fu-Yao Liu, Chih-Hung Chang
2019 Sensors and materials  
Yao-Hong  ... 
doi:10.18494/sam.2019.2254 fatcat:dadcxfamgzc6pk7os3zzvgvzca

Atrial Fibrillation in Primary Aldosteronism

Chien-Ting Pan, Cheng-Hsuan Tsai, Zheng-Wei Chen, Yi-Yao Chang, Vin-Cent Wu, Chi-Sheng Hung, Yen-Hung Lin, the TAIPAI Study Group
2020 Hormone and Metabolic Research  
AbstractPrimary aldosteronism (PA) is the most common cause of secondary hypertension. Increasing evidence has demonstrated an increased cardiovascular risk in patients with PA compared to those with essential hypertension (EH), including atrial fibrillation (AF), the most prevalent arrhythmia among adults that is associated with an elevated risk of subsequent cerebro-cardiovascular adverse events. The mechanisms of increased prevalence of AF in PA patients are complex. Excessive aldosterone
more » ... sive aldosterone production is regarded to be a key component in the pathogenesis of AF, in addition to arterial hypertension and electrolyte imbalance. In addition, several translational and clinical studies have reported that structural remodeling with atrial fibrosis and electrical remodeling with arrhythmogenicity induced by an excess of aldosterone also play major roles in AF genesis. Clinical studies from several registries and meta-analysis have reported an increased prevalence and risk of AF in PA patients compared to EH patients. Recent trials have further demonstrated a reduction in the risk of new-onset atrial fibrillation (NOAF) after adrenalectomy, while the results of medical treatment with mineralocorticoid receptor antagonists (MRAs) have been inconsistent. This review outlines the current evidence of the relationship between PA and AF, and highlights recent progress in the management of PA with regards to the development of AF.
doi:10.1055/a-1141-5989 pmid:32289838 fatcat:742nfhilsnanfnuzytojzje4bu

Eosinophilic meningitis caused byAngiostrongylus cantonensis: report of 17 cases

Hung-Chin Tsai, Yung-Ching Liu, Calvin M Kunin, Susan Shin-Jung Lee, Yao-Shen Chen, Hsi-Hsun Lin, Tsung-Hung Tsai, Wei-Ru Lin, Chun-Kai Huang, Muh-Yong Yen, Chuan-Min Yen
2001 American Journal of Medicine  
Figure 3 . 3 Changes in mean serum immunoglobulin E levels during the 6-month follow-up after the 1999 epidemic.Eosinophilic Meningitis Caused by Angiostrongylus Cantonensis/Tsai et al Punyagupta et  ... 
doi:10.1016/s0002-9343(01)00766-5 pmid:11498063 fatcat:kghiad72gjgk5pqxos22teirem

Rare Orbital Metastasis Originating from Ampullary Adenocarcinoma

Yung-En Tsai, Ke-Hung Chien, Yao-Feng Li, Shiue-Wei Lai
2021 Medicina  
Orbital metastasis from ampullary carcinoma is rare, with no previously reported cases. Case presentation: We report the case of a 60-year-old man who complained of a right-sided headache, blurred vision, progressive proptosis, ptosis, and right eye pain for 3 months. His past medical history included an ampullary adenocarcinoma stage IIIA treated via the Whipple procedure and adjuvant chemoradiotherapy 1 year ago. However, he was lost to follow-up. Computed tomography of the orbit showed a
more » ... orbit showed a soft tissue lesion in the right orbital fossa measuring 3.3 × 2 × 2 cm. An orbital mass biopsy demonstrated an intestinal-type adenocarcinoma that tested positive for cytokeratins 7 and 20 and CDX2 on immunohistochemical staining. The pathologic diagnosis was metastatic adenocarcinoma from the ampulla of Vater. Despite oncological treatment, the patient's illness progressed. He received palliative treatment and died 1 month later. Conclusions: We presented a rare case of orbital metastasis from ampullary adenocarcinoma. This should be considered in the differential diagnosis of patients with a history of ampullary adenocarcinoma who present with symptoms referring to the relevant locations.
doi:10.3390/medicina57111238 pmid:34833456 pmcid:PMC8624159 fatcat:36eu42mfgzgltfd2sfgvhiw6ea

Post Selection Inference with Incomplete Maximum Mean Discrepancy Estimator [article]

Makoto Yamada, Denny Wu, Yao-Hung Hubert Tsai, Ichiro Takeuchi, Ruslan Salakhutdinov, Kenji Fukumizu
2018 arXiv   pre-print
Measuring divergence between two distributions is essential in machine learning and statistics and has various applications including binary classification, change point detection, and two-sample test. Furthermore, in the era of big data, designing divergence measure that is interpretable and can handle high-dimensional and complex data becomes extremely important. In the paper, we propose a post selection inference (PSI) framework for divergence measure, which can select a set of statistically
more » ... et of statistically significant features that discriminate two distributions. Specifically, we employ an additive variant of maximum mean discrepancy (MMD) for features and introduce a general hypothesis test for PSI. A novel MMD estimator using the incomplete U-statistics, which has an asymptotically Normal distribution (under mild assumptions) and gives high detection power in PSI, is also proposed and analyzed theoretically. Through synthetic and real-world feature selection experiments, we show that the proposed framework can successfully detect statistically significant features. Last, we propose a sample selection framework for analyzing different members in the Generative Adversarial Networks (GANs) family.
arXiv:1802.06226v1 fatcat:kj7sqpvqkreexo7unkkkttclaa

Improving One-Shot Learning through Fusing Side Information [article]

Yao-Hung Hubert Tsai, Ruslan Salakhutdinov
2018 arXiv   pre-print
Supplementary for Improving One-Shot Learning through Fusing Side Information Yao-Hung Hubert Tsai † Ruslan Salakhutdinov † † School of Computer Science, Machine Learning Department, Carnegie Mellon University  ...  with ReViSE (Tsai et al., 2017) .  ... 
arXiv:1710.08347v2 fatcat:cm5chbpui5fc7kfxl5ffcjnncm

Left ventricular remodeling and dysfunction in primary aldosteronism

Cheng-Hsuan Tsai, Chien-Ting Pan, Yi-Yao Chang, Zheng-Wei Chen, Vin-Cent Wu, Chi-Sheng Hung, Yen-Hung Lin
2020 Journal of Human Hypertension  
Liao et al. (2015) [134] LVEF APA - Adrenalectomy LVEF showed no significant change after adrenalectomy Liao et al. (2015) [74] LVEF PA EH - No difference between PA and EH Hung  ...  Cesari et al. (2016) [71] Yang et al. (2017) [72] LVEF PA EH - No difference between PA and EH Hung et al. (2017) [101] LVEF PA EH - No difference between PA and EH Chen et al.  ... 
doi:10.1038/s41371-020-00426-y pmid:33067554 fatcat:2p3lr5ukmfanxi7yxxfadrysdm
« Previous Showing results 1 — 15 out of 1,720 results