cv

Postdoctoral Researcher, Machine Learning

University of Tübingen / Donders Institute

• Conducted research on Unsupervised 3D Representation Learning for robust Anomaly Detection in Medical Imaging.
• Improved psychosis detection by 6.7% AUROC through Self-Supervised Masked Surface Vision Transformer.
• Researched on Multi-Modal Foundation Model for Neurodegenerative Disease Detection.

Senior Machine Learning Scientist, Domain Transfer and Privacy

National Health Research Institutes

• Introduced Privacy-aware (FedCM) to enhance heterogeneity in Federated Learning for Multi-Source Alzheimer’s Detection.
• Identified and differentiated key Regions of Interest (ROI) related to normal brain degeneration and Alzheimer’s disease.
• Mentored students on robust multi-site ADHD detection using Conditional-adversarial Domain Adaptation Network (CDAN).

Doctoral Researcher, Affective Electrophysiology Responses under Audio/Visual Stimuli

Behavioral Informatics & Interaction Computation Lab (BIIC)

• Enhanced physiological emotion recognition by 3.5% UAR through Transformer-based Personalized modeling.
• Achieved an 8% UAR increase in personality recognition using Multi-Head Attention Graph Convolutional Network (GCN).
• Conducted in-depth studies on subjective/intended emotion reactions based on ECG/EDA/EEG signals with Shapley analysis.

Machine Learning Scientist, Neurocognitive Assessment using Brain MRI

National Health Research Institutes

• Spearheaded MRI research pipeline: Data-Preprocessing (FreeSurfer, SPM), Modeling, and Explainable AI tools (GradCAM++).
• Increased 8.98% UAR for 3-class Alzheimer’s detection using 3D Autoencoder with multi-view contrastive loss (BSEN).
• Attained SOTA UAR of 80.2% in binary Face Processing Ability classification using Event-Contrastive Node2Vec Embedding.

PhD

National Tsing Hua University (NTHU)

Electrical Engineering

• Got President Scholarship which is given to promising students in EE Dept.

B.S.

National Tsing Hua University (NTHU)

Electrical Engineering

C-Media Electronics, Inc.

Senior Machine Learning Scientist

• Led a team of 5 developing a Real-Time Fully-Convolutional Speech De-reverberation engine with SRMR score 4.774.
• Oversaw the entire R&D pipeline: Data Collection/Synthesis, Model/Metric Design, Error Analysis, and Model Compression.
• Pioneered the development of speech cloning from previously unseen sources using Generative Adversarial Networks (GAN).

Inventec AI Center

Machine Learning Scientist, Cardiovascular Disease Detection

• Led the development of End-to-End Heart Disease Anomaly Detection system with Time-Series Multi-Lead ECG.
• Proposed Mixed-Domain Self-Attention Resnet (MDARsn) to handle missing data and outlier ECG for robust Multi-Label Detection.
• Received the Best Challenge Poster award in the PhysioNet/CinC Challenge 2021 and secured a US patent (US20230153575A1).

Institute for Information Industry (III)

Senior Machine Learning Engineer

• Led a team of 5 developing a deep video retrieval system to accelerate fake news screening.
• Proposed a novel video retrieval system with 95.1% retrieval precision and orchestrated with Docker.
• The system is adopted by two prominent NGO fake news checkers, Taiwan FactCheck Center and MyGoPen.

Gamania Digital Entertainment Co., Ltd.

Machine Learning Engineer

• Developed Multi-modal behavior profiling system utilizing Speech, Face, and Gesture for AI-driven hiring recommendations.
• Designed and implemented a real-time, Multi-person Multi-modal data collection system to complement the profiling system.

Program Committee Member

IEEE ICKII, IEEE ICEIB, IARIA CENTRIC

Served on the program committee for IEEE ICKII, IEEE ICEIB, and IARIA CENTRIC.

Best Challenge Poster

PhysioNet/CinC Challenge, Computing in Cardiology Society

Awarded the Best Challenge Poster in the PhysioNet/CinC Challenge 2021.

Travel Grants

IEEE Signal Processing Society

Received travel grants for ICASSP 2019.

Paper Presentation Scholarship

ACLCLP, Taiwan

Scholarship for presenting research at ACLCLP.

AI Scholarship

Adbertech Inc., Taiwan

President Scholarship

National Tsing Hua University

Scholarship given to promising new PhD students in the EE Department.

Summer Academic Exchange Program

Fudan University, China

A Media-Guided Attentive Graphical Network for Personality Recognition Using Physiology

IEEE Transactions on Affective Computing

Journal paper on personality recognition from physiological signals.

A Mixed-Domain Self-Attention Network for Multilabel Cardiac Irregularity Classification Using Reduced-Lead Electrocardiogram

Computing in Cardiology (CinC) 2021, Brno, Czech Republic

Federated Learning via Conditioned Mutual Learning for Alzheimer Disease Classification on T1w MRI

43rd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC) 2021

Cross Corpus Physiological-based Emotion Recognition Using a Learnable Visual Semantic Graph Convolutional Network

MM'20: The 28th ACM International Conference on Multimedia (ACMMM) 2020

Conditional Domain Adversarial Transfer for Robust Cross-Site ADHD Classification Using Functional MRI

ICASSP 2020, Barcelona, Spain

A Siamese Content-Attentive Graph Convolutional Network for Personality Recognition Using Physiology

ICASSP 2020, Barcelona, Spain

Behavior Score-Embedded Brain Encoder Network for Improved Classification of Alzheimer Disease Using Resting State fMRI

EMBC 2020, Montreal, QC, Canada

An Attribute-invariant Variational Learning for Emotion Recognition Using Physiology

ICASSP 2019, Brighton, United Kingdom

An Event-contrastive Connectome Network for Automatic Assessment of Individual Face Processing and Memory Ability

ICASSP 2019, Brighton, United Kingdom

Annotation Matters: A Comprehensive Study on Recognizing Intended, Self-reported, and Observed Emotion Labels using Physiology

8th International Conference on Affective Computing and Intelligent Interaction (ACII) 2019, Cambridge, United Kingdom

Integrating Perceivers Neural-Perceptual Responses Using a Deep Voting Fusion Network for Automatic Vocal Emotion Decoding

ICASSP 2018, Calgary, AB, Canada

A Triplet-Loss Embedded Deep Regressor Network for Estimating Blood Pressure Changes Using Prosodic Features

ICASSP 2018, Calgary, AB, Canada