Haiyan Xu

About

I am a Ph.D. candidate in the Communications Group at the Edward S. Rogers Sr. Department of Electrical and Computer Engineering, University of Toronto. My research centres on EEG-based emotion recognition, with a focus on cross-subject generalization through graph neural networks, domain adaptation, and semi-supervised learning. I also study how benchmark protocol choices — label formats, train/test splits, preprocessing pipelines — shape reported accuracy in affective computing.

I hold a B.A.Sc. (Honours, Minor in Bio-Engineering) and an M.A.Sc. in Electrical and Computer Engineering, both from the University of Toronto.

Research Interests

• EEG-based emotion recognition
• Graph neural networks for biosignals
• Domain adaptation and cross-subject transfer learning
• Semi-supervised learning
• Benchmark methodology and reproducibility in affective computing

My current work evaluates how label type (stimulus-based vs. self-report), dataset protocol, and preprocessing choices affect the reported performance of EEG emotion models across DEAP, FACED, SEED, and DREAMER.

Publications

1. Affective States Classification using EEG and Semi-supervised Deep Learning Approaches
   H. Xu, K. N. Plataniotis
   IEEE 18th International Workshop on Multimedia Signal Processing (MMSP), 2016
2. EEG-Based Affect States Classification using Deep Belief Networks
   H. Xu, K. N. Plataniotis
   IEEE Digital Media Industry & Academic Forum (DMIAF), 2016
3. Subject Independent Affective States Classification Using EEG Signals
   H. Xu, K. N. Plataniotis
   IEEE Global Conference on Signal and Information Processing (GlobalSIP), 2015
4. Application of Multivariate Empirical Mode Decomposition in EEG Signals for Subject Independent Affective States Classification
   H. Xu, K. N. Plataniotis
   19th International Conference on Circuits, Systems, Communications and Computers (CSCC), 2015
5. Affect Recognition Using EEG Signal
   H. Xu, K. N. Plataniotis
   IEEE International Workshop on Multimedia Signal Processing (MMSP), Banff, Canada, 2012
6. A Real-Time Portable System for Classification of Meditation States Using EEG Signals
   H. Xu, M. S. Mhanta, C. Aimone, K. N. Plataniotis
   IEEE International Workshop on Multimedia Signal Processing (MMSP), Banff, Canada, 2012

Thesis

• Towards Automated Recognition of Human Emotions Using EEG
  H. Xu
  M.A.Sc. Thesis, Department of Electrical and Computer Engineering, University of Toronto, 2012

Experience

• MITACS Accelerate PhD Fellow · Qualcomm Canada 2013 – 2016
  EEG-based cognitive interfaces for ubiquitous applications: integrating wireless EEG with smartphone systems for real-time emotional and cognitive state recognition.
• MITACS Accelerate Intern · Qualcomm Canada 2012 – 2013
  Multi-modal affective signal processing system for real-time inference of cognitive-affective mental states from facial video and EEG.
• R&D Intern · InteraXon, Canada 2012
  Real-time portable classification of meditation states using wireless EEG.
• Summer Research Assistant (NSERC USRA) · University of Toronto 2009
  Face recognition using 3D facial images: algorithm development for subject identification and emotion classification.
• Research Assistant · University of Toronto 2008 – 2009
  2D and 3D shape-based biometric face recognition systems.

Awards

• MITACS Accelerate PhD Fellowship ($90,000 CAD) — 2013 – 2016
• MITACS Accelerate Internship ($15,000 CAD) — 2012 – 2013
• NSERC Undergraduate Student Research Award (USRA) — 2009
• Dean’s List, Electrical Engineering — 2006 – 2010 (top 8% of program)
• Best Podium Presenter, Engineering Undergraduate Research Day (UNERD), University of Toronto — 2009
• Best Design Project, ECE298 — 2007

Contact