黃彥華

Single-Cell GRN-Guided Transcription Factor Activity for Interpretable Survival Prediction and Regulatory Analysis in Cancer

生物醫學資訊研究所

單細胞多體學、基因調控網路、轉錄因子活性、CIBERSORTx、存活預測、胰臟導管腺癌、黑色素瘤、single-cell multiome、gene regulatory network、transcription factor activity、survival prediction、pancreatic ductal adenocarcinoma、melanoma

Whole-Slide Semantic Segmentation for Lung Adenocarcinoma Histology Subtyping

生物醫學資訊研究所・共同指導：梁恭豪

肺腺癌、生長模式、全切片影像、數位病理學、深度學習、語意分割、Lung Adenocarcinoma、Histological Growth Patterns、Whole Slide Image、Digital Pathology、Deep Learning、Semantic Segmentation

A Pathway-Based Trimodal Deep Learning Framework with Large Language Model Mutation Embeddings for Cancer Survival Prediction

生物醫學資訊研究所

存活分析、多模態學習、大型語言模型嵌入、持續性腫瘤突變負荷、路徑結構編碼器、注意力融合機制、Survival Analysis、Multimodal Learning、Large Language Model Embeddings、Persistent Tumor Mutational Burden、Pathway-Structured Encoder、Attention Fusion Mechanism

Enhancing Foundation Models for Histopathology via Multi-Scale Tile Assembly and Multi-Branch Encoders for Survival and Mutation Prediction

生物醫學資訊研究所

基因突變、生存分析、基礎模型、病理影像、多倍率模型、注意力機制、Gene Mutation、Survival Analysis、Foundation Model、Whole Slide Image、Multi-Scale Model、Attention Mechanism

A Multimodal Kidney Failure Prediction Model Integrating Ultrasound Imaging and Clinical Data: Toward Enhancing Image-Driven Predictive Capability

生物醫學資訊研究所・共同指導：梁恭豪

多模態學習、自編碼器、慢性腎臟病、腎臟超音波、存活分析、特徵提取、遷移學習、Autoencoder、Chronic Kidney Disease、Kidney Ultrasound、Survival Analysis、Feature Extraction、Transfer Learning、Multimodal Learning

AI-based Renal Function Prediction from Ultrasound Images via Self-Supervised Pretraining

生物醫學資訊研究所・共同指導：梁恭豪

Multimodal model、Survival analysis、self-supervised pre-training、regression/classification model、Kidney ultrasound imaging、多模態模型、存活分析、自監督式預訓練、回歸/分類模型、eGFR、腎臟超音波影像

A Novel Explainable AI-Based Framework for Cancer Prognostic Biomarkers Discovery and Precision Medicine

生物醫學資訊研究所

機器學習、可解釋人工智慧、腎臟癌、預後生物標記、癌症期別分類、Machine learning、XAI、Renal cell carcinoma、prognostic biomarkers、stage classification

Deep Learning for Prokaryotic Transcription Start Site Prediction with Object Detection

生物醫學資訊研究所

σ70、原核生物、轉錄起始點、啟動子、prokaryotic transcription start site、promoter

Identification of potential pathogenic gene variants for familial nyctalopia using whole exome sequencing analysis

生物醫學資訊研究所

全外顯子定序分析、圖卷積網路、候選基因排序、色素性視網膜炎、夜盲症、WES、GCN、Candidate gene prioritization、Retinitis pigmentosa、Night blindness

Semantic Segmentation of Renal Ultrasound Images Using Fully Supervised and Semi-Supervised Deep Learning Models

生物醫學資訊研究所

深度學習、語義分割、腎臟超音波、監督式學習、半監督式學習、deep learning、semantic segmentation、renal ultrasound、supervised learning、semi-supervised learning

Dysregulated mechanisms associated with the suppressed immunity of CGL2 patients

生物醫學資訊研究所

RNA定序、單細胞RNA定序、先天性全身脂肪失養症、反卷積、生物網路、深度學習、Bulk RNA-seq、Single cell RNA-seq、Congenital generalized lipodystrophy、Deconvolution、Biological network、Deep learning

Deep learning to predict tumor-specific features by whole-slide images (WSI) of cancers

生物醫學資訊研究所

深度學習、數位病理切片影像、腫瘤突變負荷量、免疫療法、弱監督式學習、多實例學習、Deep learning、Whole slide image、Tumor mutational burden、Immunotherapy、Weakly supervised learning、Multiple instance learning

Application of deep learning in single cell mRNA sequencing data analysis

生物醫學資訊研究所

微陣列、單細胞 RNA 定序、循環生成對抗網路、圖捲積網路、單細胞分類、Microarray、scRNA-seq、cycleGAN、GCN、Classification of scRNA-seq

Classification stroke with head computed tomography by using convolutional neural networks

生物醫學資訊研究所

中風、電腦斷層掃描、窗幅與窗值、卷積神經網路、壓縮與激發區塊網路、弱 監督式卷積神經網路、Stroke、Computer tomography、CT windows、Convolutional neural networks、Squeeze and excitation block、WSCNet

Assessment of imputation methods for single-cell RNAseq data to analyze cell identity and respective regulatory networks

生物醫學資訊研究所

評估、基因表現量、細胞種類、插補、深度學習、機器學習、Assessment、scRNA-seq、Dropout events、Imputation、Deep learning、Machine learning

Transcriptome profiling reveals the underlying mechanisms of stress responses of a high-virulence and drug-resistance Klebsiella pneumoniae strain

生物醫學資訊研究所

克雷伯氏肺炎桿菌、次世代定序、Klebsiella pneumoniae、NGS

Using Non-negative matrix factorization to distinguish tumor- and storma- specific gene expression

生物醫學資訊研究所

癌症、免疫細胞、微陣列、非負矩陣分解法、基因表現、存活分析、cancer、immune cell、microarray、Non-negative matrix factorization、gene expression、survival analysis

Repurposing small-molecular drugs to block the interaction between PD-1 and PD-L1 using bioinformatic approaches

生物醫學資訊研究所

PD-1、docking、結構生物學、connectivity map、HBPLUS、Structural biology