曾新穆

Mixture of Decision-Focused Trading Experts with Adaptive Routing for Mean Variance Portfolio Optimization.

數據科學與工程研究所・共同指導：高竹嵐

投資組合管理、交易策略、均值變異數最佳化、目標專注學習、混和專家模型、自適應路由器、Portfolio Management、Mean-Variance Optimization、Decision-focused Learning、Mixture-of-Experts、Adaptive Routing

FedSGAD: Unsupervised Federated Cross-Modality Medical Image Translation via Structurally-Guided Adversarial Diffusion

資訊科學與工程研究所

聯邦學習、跨模態醫學影像轉換、擴散模型、醫學影像合成、個人化聯邦學習、非監督學習、Federated Learning、Cross-Modality Medical Image Translation、Diffusion Models、Medical Image Synthesis、Personalized Federated Learning、Unsupervised Learning

SleepHAS: A Multimodal-Based Framework for Sleep Stage Classification by Home-Acquirable Signals

智能系統研究所・共同指導：英家慶

睡眠階段分類、階層式多模態融合、光體積變化描記圖、呼吸訊號、活動訊號、Sleep Stage Classification、Hierarchical Multimodal Fusion、Photoplethysmography、Respiratory Signals、Activity Signals

PPAL: Prototypical Precursor-Aware Learning for Multivariate Time-Series Anomaly Prediction

數據科學與工程研究所

異常預測、多變量時間序列、預防性維護、原型學習、非監督式學習、Anomaly prediction、Preventive maintenance、Multivariate time series、Prototype learning、Unsupervised learning

LSF-SR: Latent Semantic Fusion for Sequential Recommendation via Conditional Variational Autoencoders

資訊科學與工程研究所・共同指導：英家慶

序列推薦、大型語言模型、標準化流、條件變分自編碼器、潛在語意融合、Sequential Recommendation、Large Language Models、Normalizing Flows、Conditional Variational Autoencoder、Latent Semantic Fusion

TREND: Trusted-Retrieval-Enhanced Diffusion Frameworkfor Medical Label-Noise Learning

數據科學與工程研究所・共同指導：高竹嵐

雜訊標籤、資料不平衡、擴散式模型、醫療影像分類、Noisy Labels、Imbalanced Data、Diffusion、Medical Image Classification

A Study on Information Gap Bridging in Wearable ECG Devices for Early Classification

數據科學與工程研究所・共同指導：洪慧念

穿戴式裝置、心電圖分類、訊號重建、知識蒸餾、及早序列分類、深度學習、Wearable Devices、ECG Classification、Signal Reconstruction、Knowledge Distillation、Deep Learning、Early Time Series Classification (ETSC)

Reconstruction-based Aanomaly Predictor for Exetremely Imbalanced Multivariate Time-series Anomaly Prediction

數據科學與工程研究所・共同指導：洪慧念

異常預測、異常偵測、不平衡、自編碼機、剩餘使用壽命預測、Anomaly prediction、Anomaly detection、Imbalance、Autoencoder、Remaining useful life prediction

Sensor Synthesis and Interaction Graph Learning for Generalizable Cross-Domain Human Activity Recognition

智能系統研究所

人類行為辨識、混合領域適應、圖神經網路、穿戴式感測器、感測器合成、Human Activity Recognition、Hybrid Domain Adaptation、Graph Neural Network、Wearable Sensors、Sensor Synthesis

Temporal-Relational Inter-Firm Evidence Retrieval for Explainable Stock Movement Prediction

資訊科學與工程研究所

可解釋性股價漲跌預測、時間與關係感知重排序、跨公司證據檢索、外溢效應、長尾股票、Explainable Stock Movement Prediction、Temporal-Relational Re-ranking、Inter-Firm Evidence Retrieval、Spillover Effects、Long-tail Stocks

MTSSRL-MD: Multi-Task Self-Supervised Representation Learning for EEG Signals Across Multiple Datasets

資訊科學與工程研究所

腦電圖、睡眠階段分類、自我監督式表徵學習、多資料集學習、多任務學習、EEG、sleep stage classification、self-supervised representation learning、multi-dataset learning、multi-task learning

DL-KDD: Dual-Lightness Knowledge Distillation for Action Recognition in the Dark

資訊科學與工程研究所

人體動作識別、暗處人體動作識別、知識蒸餾、偏差校正、影片分類、Human Action Recognition、Action Recognition in the Dark、Knowledge Distillation、Bias Correction、Video Classification

Transformer-based Early Classification for Diverse and Varied-length Time Series

數據科學與工程研究所

時間序列、早期分類、深度強化學習、時間空間整合、注意力機制、可變長度序列、Time Series、Early Classification、Deep Reinforcement Learning、Spatial-Temporal Integration、Attention Mechanism、Variable-Length Sequences

To predict the severity of coronary artery stenosis in coronary artery disease patients using an inverse problem algorithm technique as an artificial intelligence-based diagnostic assessment

生醫科學與工程博士學位學程

逆運算疊代法技術、SYNTAX 分數、冠狀動脈疾病、計算分析、人工智慧、Inverse Problem Algorithm、SYNTAX、Cardiovascular Artery Disease、Computational Analysis、Artificial Intelligence

MIND - Mining INsightful Patterns of Dementia from Large-Scale Health Databases for Early Assessment

智能系統研究所

電子健康記錄（EHR）、失智症風險預測、去噪技術、序列模式挖掘、診斷時間（TTD）、風險因子分類、Electronic Health Records (EHR)、dementia risk prediction、denoising techniques、sequential pattern mining、Time-to-Diagnosis (TTD)、risk factor categorization

Multi-Agent Transformer-Based Automated Imbalanced Time Series Classification with Hyperparameter Optimization

智能系統研究所

時間序列、不平衡分類、自動化機器學習、多智能體強化學習、超參數最佳化、Time Series、Imbalanced Classification、Automated Machine Learning、Multi-Agent Reinforcement Learning、Hyperparameter Optimization

Traffic Origin-Destination Demand Prediction Based on Semantic Locations

多媒體工程研究所・共同指導：王豐堅

交通預測、起訖需求、語意位置、興趣點、traffic prediction、origin-destination demand、semantic locations、points of interest

Next Week Travel Time Prediction: An Attention-Based Learning Framework Incorporating Weather and Temporal Influences

數據科學與工程研究所

長期交通預測、時間序列、深度學習、多頭注意力機制、交通時間預測、Long-term traffic prediction、Time series、Deep learning、Multi-head attention mechanism、Traffic time prediction

M3: A Multi-Agent Reinforcement Learning Framework for Integrating Multi-Granularity Financial Indicators with Multi-Incorporated Reward Function

數據科學與工程研究所

投資組合管理、強化學習、計算金融、交易策略、深度學習、portfolio management、reinforcement learning、computational finance、trading strategy、deep learning

DiPPSI : Diffusion-based Pulsative Physiological Signal Imputation

數據科學與工程研究所・共同指導：洪慧念

行動健康、三軸加速度計、條件擴散模型、信號補值、脈衝模式、mHealth、accelerometers、diffusion model、signal imputation、pulse pattern

COLA: Causally Oriented Learnable Augmentation for Multi-Domain Long-Tailed Image Classification

資訊科學與工程研究所

長尾學習、多域學習、因果關係、可學習資料增強、Long-Tailed Learning、Multi-Domain Learning、Causality、Learnable Augmentation

Hardness-aware Adversarial Negative Sampling for Recommendation

資訊科學與工程研究所

協同過濾、隱式回饋、負取樣、對抗式網路、推薦系統、Collaborative Filtering、Implicit Feedback、Negative Sampling、Adversarial Networks、Recommendation System

A Study on Hierarchy-Aware Classification Methodologies with Applications in Medical Image Recognition

資訊科學與工程研究所

階層感知分類、對比學習、原型學習、開集識別、輕量化模型、皮膚病灶識別、Hierarchy-Aware Classification、Contrastive Learning、Prototype Learning、Open-Set Recognition、Lightweight Model、Skin Lesion Recognition

A Study on Transformer-based Methodologies for Travel Time Prediction

資訊科學與工程研究所・共同指導：彭文孝

交通時間預測、智慧交通系統、長期預測、週期交通數據、Transformer、Travel time prediction、Intelligent Transportation Systems、Long-term prediction、Periodic traffc data

Prediction of Next Week's Significant Stock Trends Based on Multi-Time Granularity Technical Indicators

數據科學與工程研究所・共同指導：高竹嵐

股票趨勢、時間粒度、技術指標、時間序列、巨量資料、Stock Trend、Time Granularity、Technical Indicators、Time Series、Big Data

Stock Price Fluctuation Prediction based on Machine Learning: A Study on Taiwan Semiconductor Industry

企業管理碩士學位學程

股價波動度預測、時間序列分析、分類方法、機器學習、深度學習、Stock Price fluctuation Prediction、Time Series Analysis、Classification、Machine Learning、Deep Learning

Cross-Domain Few-Shot Sensor-Based Activity Recognition with Pseudo-Domain Labeling and Feature Learning

資訊科學與工程研究所

跨域性、人類活動識別、時間序列、小樣本學習、Cross-Domain、Few-Shot Learning、Time Series、Human Activity Recog- nition

Self-supervised Pulse-Aware and Interpretable Disentangled ECG Representation Learning

資訊科學與工程研究所

心電圖、非監督式表徵學習、自監督式學習、槽注意力模型、Electrocardiography、unsupervised representation learning、self-supervised learning、slot attention

Next-day Travel Time Estimation Using Self-Attention and Feature Fusion Learning

資訊科學與工程研究所

巨量資料、未來日旅行時間預測、時間序列、深度學習、自注意力機制、Big Data、Next-day Travel Time Estimation、Time series、Deep Learning、Self-attention

Privacy-Preserving Attention-Weighted Multi-Source Domain Adaptation for EEG Motor Imagery

數據科學與工程研究所・共同指導：洪慧念

動作意象任務、多源域權重、隱私保護、非監督式域適應、注意力機制、Motor imagery、Weighted multi-source、Privacy-preserving、Unsupervised domain adaptation、Attention mechanism

A Study of Medical Early Warning Systems Using Time-series Data

資訊管理研究所・共同指導：黃興進

醫療早期警示系統、卷積神經網絡、深度強化學習、心血管疾病分類、敗血症早期預測、Medical Early Warning System、Convolutional Neural Network、Deep Reinforcement Learning、Cardiovascular Disease Classification、Sepsis Early Prediction

Early Multi-label Classification for Long-length and Varied-length Time Series

數據科學與工程研究所・共同指導：洪慧念

早期分類、多標籤分類、時間序列分類、深度強化學習、注意力機制、Early Classification、Multi-label Classification、Time Series Classification、Deep Reinforcement Learning、Attention Mechanism

Multi-View Human Action Recognition on Incomplete Viewpoint

資訊科學與工程研究所

人體動作識別、多視角、不完全視角、知識蒸餾轉換器、Human Action Recognition、Multi-View、Incomplete Viewpoint、Knowledge Distillation Transformer

Class-dependent Automatic Data Augmentation based on Adaptive Policies for Time Series

資訊科學與工程研究所

數據增強、自適應數據增強、時間序列、心電圖、類別相關偏誤、Data Augmentation、Automatic Data Augmentation、Time series、ECG signals、Class-dependent Bias

Deep Learning in Physiological Signal Processing and Clinical Prediction: Applications in Cardiology, Critical Care, and Emergency Medicine

生醫科學與工程博士學位學程

心律不整、卷積神經網路、深度學習、心電圖、經皮冠狀動脈介入治療、ST段上升型急性心肌梗塞、心室顫動、Cardiac arrhythmias、Convolutional neural network、Deep learning、Electrocardiography、Percutaneous coronary intervention、STEMI、Ventricular fibrillation

Integrated Interictal Epileptiform Discharges Detection from Scalp EEG Signals: Methods and Evaluation

數據科學與工程研究所・共同指導：洪慧念

癲癇波偵測、頭皮腦波圖、深度學習、規則法、模板比對、Iterictal Epileptiform Discharges Detection、Scalp Electroencephalography、Deep Learning、Rule-based、Template Matching

Multi-Step Predictions for Commodity Prices with High-Dimensional Mixed-Frequency Exogenous Information

數據科學與工程研究所・共同指導：洪慧念

原物料價格預測、經濟指標、巨量資料、混和頻率時間序列、深度學習、多步預測、commodity prices prediction、economic indicators、big data、mixed-frequency time series、neural network、multi-step forecasting

Continuous Spatial-Temporal Forecasting with Multi-Objective Optimization for Traffic Demand Prediction

資訊科學與工程研究所

時間序列、時空資料、長期時空預測、帕累托最佳化、多目標最佳化、出租車需求預測、Spatial and temporal data、Long-term spatial-temporal prediction、Pareto optimization、Multi-objective optimization、Taxi demand prediction

Semi-Supervised Learning for Long-Tailed Recognition

數據科學與工程研究所

巨量資料、半監督式學習、長尾識別、影像分類、卷積神經網路、big data、semi-supervised learning、long-tailed recognition、image classification、convolutional neural network

Modeling and Prediction on Effects of Anomaly Events: Application on Traffic Congestion Prediction for Incidents

資訊科學與工程研究所

異常事件影響性、巨量資料、時間序列、壅塞影響預測、注意力機制、深度學習、Effects of Anomaly Event、Big Data、Time series、Traffic Congestion Prediction、Attention Mechanism、Deep Learning

Early Time Series Anomaly Prediction with Multi-Objective Optimization

資訊科學與工程研究所・共同指導：煙淦

時間序列、異常預測、多目標優化、早期時間序列異常預測、預測性維護、Predictive maintenance、Anomaly prediction、Multi-objective optimization、Early time series anomaly prediction

A Study on Early Time Series Classification Methodologies with Applications in Cardiovascular Disease Detection

資訊科學與工程研究所・共同指導：煙淦

時間序列早期分類、深度類神經網路、多目標優化演算法、深度強化學習、心血管疾病偵測、Ealy Time Series Classification、Deep Neural Networks、Multi-objective Optimization、Deep Reinforcement Learning、Cardiovascular Disease Classification

Generalized Machine Learning-based Framework for Risk Prediction in ICU

數據科學與工程研究所

風險預測、機器學習、多模態學習、臨床時間序列、心電圖、risk prediction、machine learning、multimodal learning、clinical time series、ECG waveform

Multi-­Task Multi-­Label Ocular Disease Classification via Weakly Supervised Localization

資訊科學與工程研究所・共同指導：陳健

眼部疾病分類、患者層級診斷、眼睛層級診斷、多任務學習、多標籤分類、弱監督式物體定位、Ocular Disease Classification、Patient-level Diagnosis、Eye-level Diagnosis、Multi-task Learning、Multi-label Classification、Weakly Supervised Object Localization

Enhancing Image Key-Point Detection Through Adaptive Composition

數據科學與工程研究所

物件偵測、關鍵點偵測、馬賽克數據增強方法、金字塔級別自適應加權、金字塔級別融合、Object Detection、Key-point Detection、Mosaic Data Augmentation、Adaptive Weighted Pyramid Levels、Pyramid Levels Aggregation

Early Time Series Classification for Anomaly Prediction

數據科學與工程研究所・共同指導：洪慧念

時間序列、早期分類、異常預測、預測性維護、Time series、Early classification、Anomaly prediction、Predictive maintenance

Hierarchical Graph Representation Learning for Origin-Destination Traffic Demand Prediction

資訊科學與工程研究所

出租車需求預測、起迄交通需求預測、圖卷積網路、層次分聚方法、Taxi demand prediction、Origin-destination demand prediction、Graph Convolution、Hierarchical clustering

Effectiveness Evaluation for Incorporating Expert Trading Behavior in Stock Prediction Models

數據科學與工程研究所・共同指導：高竹嵐

股價趨勢預測、機器學習、時間序列、專家投資行為、Stock Trend Prediction、Machine Learning、Time Series、Expert Trading Behavior

Next-Basket Recommendation with Serendipity

資訊科學與工程研究所

購物籃推薦、驚喜推薦、協同過濾、使用者行為分析、日常生活購、Next-basket Recommendation、Serendipitous Recommendation、Collaborative Filtering、User Behavior Mining、Grocery Shopping

Deep Learning-Based Discrete-Time Survival Prediction for Non-Small Cell Lung Cancer

數據科學與工程研究所

存活分析、深度學習、醫學影像、表徵學習、非小細胞肺癌、Survival analysis、Deep learning、Medical image、Feature Representation、Non-Small Cell Lung Cancer

A Study on Trend Analysis in High Utility Itemset Mining

電機資訊國際學程

資料探勘、高效益項目集、趨勢分析、效益型樣探勘、交易型資料庫、時間型資料庫、資料串流、項目集探勘、Data Mining、High UtilityItemsets、Trend Analysis、Utility Pattern Mining、Transaction Database、Temporal Database、Data Stream、Itemset Mining

A Deep Learning-based Approach for Label Noise-Resistant Classification on Medical Images - A Case Study on Diabetic Retinopathy

資訊科學與工程研究所

標籤噪音、深度學習、醫學影像分類問題、糖尿病視網膜病變、label noise、deep neural network、medical data classification、Diabetic Retinopathy

Video Content-based Action Sequence Forecasting using Hierarchical Recurrent Neural Networt

資訊科學與工程研究所・共同指導：陳健

動作序列預測、階層式、循環神經網路、注意力機制、深度學習、Action Sequence Forecasting、Hierarchical、Recurrent Neural Network、Attention Mechanism、Deep Learning

Enhancing Stock Trend Prediction Models by Mining Relational Graphs of Stock Prices

資訊科學與工程研究所・共同指導：陳健

股市預測、時間序列、知識表示學習模型、動態時間規整、Stock trend prediction、Time series、Knowledge graph embedding model、Dynamic time warping

Boosting Multi-task Learning Through Combination of Task Labels - with Applications in ECG Phenotyping

數據科學與工程研究所

多目標學習、輔助任務、不平衡資料、深度學習、心電圖疾病預測、Multi-task Learning、Auxiliary Task、Imbalanced Data、Deep Learning、Electrocardiogram Phenotyping

Supervised Multimodal Contrastive Coding for Correcting Imperfect Temporal Modalities with Applications in Sentiment Analysis

數據科學與工程研究所・共同指導：嚴力行

多模態機器學習、多模態情感分析、不完整模態、深度學習、多模態對比學習、Multimodal Machine Learning、Multimodal Sentiment Analysis、Imperfect Modality、Deep Learning、Multimodal Contrastive Learning

Continual Learning with Attentive Recurrent Neural Networks for Sequential Data Classification

資訊科學與工程研究所

持續性學習、循環神經網路、機器學習、任務增加學習、筆跡分類、人類行為識別、continual learning、recurrent neural network、machine learning、task incremental learning、stroke data classification、human activity recognition

Deep Metric Learning for Unsupervised Electrocardiogram Representation and Phenotyping

數據科學與工程研究所・共同指導：劉建良

心電圖分析、深度學習、無監督式學習、特徵學習、度量學習、Electrocardiogram analysis、deep neural network、unsupervised learning、representation learning、metric learning

Efficient and Light-Weight Object Detection Techniques for Edge Computing

資訊學院資訊學程

物件偵測、輕量化模型、邊際運算、模型壓縮、Object Detection、Model Compression、Edge Computing、Light-Weight Model

A Deep Learning-Based Lightweight Approach for User Identity Recognition on Mobile Devices

資訊科學與工程研究所・共同指導：陳健

身份識別、卷積神經網路、輕量化模型、生物識別分析、Identity Recognition、Convolutional Neural Networks、Lightweight Model、Biometric Analysis

Heterogeneous Information Fusion Method for Significant Financial Trend Prediction

資訊科學與工程研究所

異質性資料整合、輿論探勘、機器學習、自然語言處理、heterogeneous information fusion、opinion mining、machine learning、natural language processing

A Deep Learning Approach for Long-term Multi-Step Citywide Crowd Flow Prediction

資訊科學與工程研究所

群體流量預測、時間序列、時空資料、⻑期預測、多步預測、深度學習、Crowd Flow Prediction、Spatial and Temporal Data、Long-term Prediction、Multi-step Architecture、Deep Learning

High Dimensional Multivariate Time Series Prediction Methods with Applications on Economic Indicator Forecasting

資訊科學與工程研究所・共同指導：陳健

巨量資料、多變量時間序列預測、深度學習、類神經網路、經濟指標預測、big data、multivariate time series forecasting、deep neural network、economic indicator prediction

Mining Early Disease Risk Patterns for Osteoarthritis from Large-Scale Clinical Databases

資訊學院資訊學程

臨床資料庫、資料探勘、病例對照研究、骨關節炎、Clinical Databases、Data Mining、Case-Control Study、Osteoarthritis

Medical Image Classification Using Deep Learning with Integration of Disease Features Detection

資訊科學與工程研究所・共同指導：盧鴻興

深度學習、機器學習、卷積神經網路、醫學影像、影像分類、物件偵測、deep learning、machine learning、convolutional neural network、medical image、image classification、object detection

Efficient Prediction of Significant Trends Using Social Media Data : Case Studies on Stock Market

資訊科學與工程研究所

趨勢預測、社群媒體、分類問題、意⾒探勘、Trend Prediction、Social Media、Classification、Opinion Mining

Long-Term Traffic Time Prediction Using Deep Learning with Integration of Weather Effect

網路工程研究所

長期預測、行車時間預測、資料不平衡問題、集成模型、Long-Term Prediction、Traffic Time Prediction、Data Imbalanced Problem、Ensemble Model

Deep Discriminative Features Learning and Sampling for Imbalanced Data Problem

資訊科學與工程研究所・共同指導：劉建良

資料不平衡問題、人造資料樣本、特徵嵌入空間、中心損失函數、三重損失函數、Imbalanced Data、Synthetic Sampling、Feature Embedding、Center Loss、Triplet Loss

Multivariate Time Series Early Classification with Interpretability using Attention Mechanism

資訊科學與工程研究所・共同指導：劉建良

時間序列之早期預測、類神經網路、注意力機制、early classification on time-series、neural network、attention

A Framework and Evaluation for Prediction of Ventricular Fibrillation Onset based on Deep Learning

資訊科學與工程研究所

心電圖分析、心室顫動、時間序列、深度學習、早期預測、ECG Analysis、Ventricular Fibrillation、Time Series、Deep Learning、Early Prediction

Accurate Diagnosis of Endoscopic Mucosal Healing in Ulcerative Colitis by Deep Learning and Machine Learning

資訊科學與工程研究所・共同指導：盧鴻興

深度學習、機器學習、轉移學習、電腦輔助診斷系統、黏膜癒合、潰瘍性結腸炎、deep learning、machine learning、transfer learning、computer aided diagnostic system、mucosal healing、ulcerative colitis

Trend Prediction on Multivariate Time Series by Using Deep Learning with Case Studies in the Stock Market

多媒體工程研究所

時間序列分析、股票價格預測、分類方法、深度學習、趨勢預測、Time Series Analysis、Stock Price Prediction、Classification、Deep Learning、Trend Prediction

Emerging Topics Prediction on Social Media

資訊科學與工程研究所

社群媒體、新興模式、主題建模、意見挖掘、串流資料挖掘、Social Media、Emerging Pattern、Topic Modeling、Opinion Mining、Data Stream Mining

Long-Term User Location Prediction Using Deep Learning and Periodic Pattern Mining

資訊科學與工程研究所

長期預測、位置預測、軌跡探勘、移動模式、Long-Term Prediction、Location Prediction、Trajectory Mining、Mobility Pattern

Multivariate Time Series Early Classification by Using Multi-Domain Deep Neuron Networks

資訊科學與工程研究所

時間序列、多變量、早期分類、深度類神經網路、資料探勘、Time Series、Multivariate、Early Classification、Deep Neural Network、Data Mining

A Framework for Time Series Trend Prediction with Applications on Material Price Prediction

資訊科學與工程研究所

時間序列分析、原物料價格預測、分類方法、深度學習、循序樣式、time series analysis、material price prediction、classification、deep learning、sequential patterns