- Sports Analytics and Performance
- Sports Performance and Training
- Antimicrobial Resistance in Staphylococcus
- Anomaly Detection Techniques and Applications
- Human Pose and Action Recognition
- Bacterial Identification and Susceptibility Testing
- Advanced Neural Network Applications
- Time Series Analysis and Forecasting
- Topic Modeling
- Advanced Vision and Imaging
- Multimodal Machine Learning Applications
- Advanced Image and Video Retrieval Techniques
- Cancer-related molecular mechanisms research
- MicroRNA in disease regulation
- Antibiotic Resistance in Bacteria
- Robotics and Sensor-Based Localization
- Misinformation and Its Impacts
- Circular RNAs in diseases
- Sentiment Analysis and Opinion Mining
- Advanced Sensor and Energy Harvesting Materials
- Bacterial biofilms and quorum sensing
- Advanced Thermoelectric Materials and Devices
- Educational Games and Gamification
- Housing Market and Economics
- Video Analysis and Summarization
National Yang Ming Chiao Tung University
2021-2025
Cornell University
2023-2025
Chung Shan Medical University
2014-2024
Chung Shan Medical University Hospital
2021-2024
Kunming University of Science and Technology
2023-2024
Jilin Medical University
2016-2024
Yanbian University
2018-2024
Ministry of Transportation and Communications
2023
Anhui University
2023
Guangzhou Medical University
2020-2021
Consider end-to-end training of a multi-modal vs. uni-modal network on task with multiple input modalities: the receives more information, so it should match or outperform its counterpart. In our experiments, however, we observe opposite: best can network. This observation is consistent across different combinations modalities and tasks benchmarks for video classifications. paper identifies two main causes this performance drop: first, networks are often prone to overfitting due increased...
Objective In this study, we evaluated the efficacy of hydroxychloroquine (HCQ) against coronavirus disease 2019 (COVID-19) via a randomized controlled trial (RCT) and retrospective study. Methods Subjects admitted to 11 designated public hospitals in Taiwan between April 1 May 31, 2020, with COVID-19 diagnosis confirmed by pharyngeal real-time RT-PCR for SARS-CoV-2, were at 2:1 ratio stratified mild or moderate illness. HCQ (400 mg twice d 200 daily 6 days) was administered. Both study...
Current state-of-the-art object detection and segmentation methods work well under the closed-world assumption. This setting assumes that list of categories is available during training deployment. However, many real-world applications require detecting or segmenting novel objects, i.e., never seen training. In this paper, we present, UVO (Unidentified Video Objects), a new benchmark for openworld class-agnostic in videos. Besides shifting focus to open-world setup, significantly larger,...
Keypoint detection & descriptors are foundational technologies for computer vision tasks like image matching, 3D reconstruction and visual odometry. Hand-engineered methods Harris corners, SIFT, HOG have been used decades; more recently, there has a trend to introduce learning in an attempt improve key-point detectors. On inspection however, the results difficult interpret; recent learning-based employ vast diversity of experimental setups design choices: empirical often reported using...
Compressible, conductive, ultralight, and superhydrophobic graphene aerogels (GAs) are promising for wearable electronics adsorption applications. However, the unsatisfactory sensing performances lack of multiscale structural regulation still impede development multifunctional GAs. Herein, a aerogel based on graphene/silk is reported─a highly ordered three-dimensional (3D) reduced oxide (rGO) conductive network established by an alkali-induced hydrothermal self-assembly strategy, while silk...
Over-expression of AdeABC efflux pump stimulated continuously by the mutated AdeRS two component system has been found to result in antimicrobial resistance, even tigecycline (TGC) multidrug-resistant Acinetobacter baumannii (MRAB). Although insertion sequence, ISAba1, contributes one mutations, detail mechanism remains unclear. In present study we collected 130 TGC-resistant isolates from 317 carbapenem resistant MRAB (MRAB-C) isolates, and 38 them were characterized with ISAba1 adeS gene....
Background Staphylococcus aureus causes a variety of severe infections such as bacteremia and sepsis. At present, 60–80% S. isolates from Taiwan are methicillin resistant (MRSA). It has been shown that certain MRSA clones circulate worldwide. The goals this study were to identify in correlate the molecular types with their phenotypes. Methods A total 157 bacteremic patients collected nine medical centers. They typed based on polymorphisms agr, SCCmec, MLST, spa, dru. Phenotypes characterized...
We aimed to explore the mechanism of pramipexole (PPX) actions in treatment Parkinson's disease (PD).Genes related PD and PPX were screened through bioinformatics retrieval.The model was constructed by applying 1-methyl-4-phenylpyridinium (MMP+).The RNA expression levels circSNCA, SNCA, apoptosisrelated genes (BCL2, CASP3, BAX, PTEN P53) miR-7 detected qRT-PCR.Protein determined western blot.The interactions between circSNCA-miR-7-SNCA verified dual luciferase assay immunofluorescence...
The aim of this study was to determine the function NEAT1/miR-23a-3p/SMC1A axis in cell proliferation and apoptosis acute myeloid leukemia (AML). Microarray analysis used screen differentially expressed lncRNAs/miRNAs/mRNAs primary AML cells. expression nuclear paraspeckle assembly transcript 1 (NEAT1), miR-23a-3p, structural maintenance chromosome alpha (SMC1A) cells THP-1 were measured by quantitative real-time polymerase chain reaction (qRT-PCR). A Cell Counting Kit-8 (CCK-8) assay...
Gliomas are common life-threatening cancers, mainly due to their aggressive nature and frequent invasiveness long non-coding RNAs (lncRNAs) emerging as promising molecular targets. Therefore, we explored the regulatory mechanisms underlying putative involvement of lncRNA PAX-interacting protein 1- antisense RNA1/ETS proto-oncogene 1/kinesin family member 14 (PAXIP1-AS1/ETS1/KIF14) axis in glioma cell invasion angiogenesis. Firstly, identified differentially expressed PAXIP1-AS1 associated...
Hepatocellular carcinoma (HCC) is the most common subtype of liver malignancy, and it characterized by poor prognosis because cancer stem cell (CSC)-mediated high postsurgical recurrence rates. Thus, targeting CSCs, or HCC cells with CSC-like properties, an effective strategy for therapy. Here, using long noncoding RNA (lncRNA) microarray analysis, we identified a novel lncRNA termed lncCAMTA1 that increased in both CSCs HCC. High expression indicates clinical outcome. In vitro vivo...
This paper presents a novel task together with new benchmark for detecting generic, taxonomy-free event boundaries that segment whole video into chunks. Conventional work in temporal segmentation and action detection focuses on localizing pre-defined categories thus does not scale to generic videos. Cognitive Science has known since last century humans consistently videos meaningful happens naturally, without being explicitly asked do so. Here, we repeat these cognitive experiments...
Abstract Objectives The recent emergence of carbapenem-resistant Enterobacterales poses a major and escalating threat to global public health. This study aimed analyse the distribution antimicrobial resistance harbouring variant OXA-48-like carbapenemase-related genes. Methods isolates were collected from Antimicrobial Testing Leadership Surveillance (ATLAS) programme during 2018–2021. Comprehensive susceptibility testing β-lactamase gene detection also conducted, along with statistical...
With the development of 5G information technology, higher requirements are imposed on heat dissipation performance conventional polymer-based electronic packaging materials. Fabrication an interconnected graphene nanosheet framework in polymers is efficient method. However, current methods for preparation frameworks typically complex and have low cross-linking densities. Herein, we prepared high-quality with high cross-link density. When content was merely 9 wt %, in-plane thermal...
Efficiently reconstructing accurate 3D models from monocular video is a key challenge in computer vision, critical for advancing applications virtual reality, robotics, and scene understanding. Existing approaches typically require pre-computed camera parameters frame-by-frame reconstruction pipelines, which are prone to error accumulation entail significant computational overhead. To address these limitations, we introduce VideoLifter, novel framework that leverages geometric priors...
ABSTRACT Monoclonal antibodies (mAbs) that specifically recognize cell surface glycans associated with cancer and infectious disease hold tremendous value for both basic research clinical applications. However, high-quality anti-glycan mAbs, especially those sufficiently high affinity specificity, remain scarce, highlighting the need protein engineering approaches based on rational design or directed evolution enable optimization of antigen-binding properties. To this end, we sought to...
Recent advancements in LLM-based multi-agent (LLM-MA) systems have shown promise, yet significant challenges remain managing communication and refinement when agents collaborate on complex tasks. In this paper, we propose \textit{Talk Structurally, Act Hierarchically (TalkHier)}, a novel framework that introduces structured protocol for context-rich exchanges hierarchical system to address issues such as incorrect outputs, falsehoods, biases. \textit{TalkHier} surpasses various types of...
Multivariate time-series forecasting is vital in various domains, e.g., economic planning and weather prediction. Deep train-from-scratch models have exhibited effective performance yet require large amounts of data, which limits real-world applicability. Recently, researchers leveraged the representation learning transferability pre-trained Large Language Models (LLMs) to handle limited non-linguistic datasets effectively. However, incorporating LLMs with data presents challenges adaptation...
To test the effects of Tanshinone IIA (Tan IIA) on cell viability, cycle, apoptosis, and autophagy human glioma U251 by regulating phosphatidylinositol 3-kinase (PI3K)/Akt/mammalian target rapamycin (mTOR) signal pathway.Tan PI3K agonist (740 Y-P) were used to treat cells U251. MTT assay was assess viability flow cytometry detect apoptosis cycle. The expressions apoptosis-related proteins (Bcl-2 Bax), autophagy-related (LC3B Beclin 1) PI3K/Akt/mTOR pathway-associated (p-PI3K, p-Akt p-mTOR)...