Yi Hao Chan

Welcome! 😄

I’m currently a postdoctoral fellow in the College of Computing and Data Science, Nanyang Technological University, Singapore. I’ve completed my undergraduate degrees (Business and Computing, with specialisations in Business Analytics, Data Science & Analytics, AI) and PhD (Computer Science, specifically in brain imaging, multimodal learning and explainable AI) in NTU as well.

My research interests are within the intersection of Machine Learning and Neuroscience. Currently, my main research focus is on Temporal Graph Interpretability. Specifically, I am interested in modelling multimodal datasets of human brain activity to identify salient spatiotemporal patterns, improving our understanding of structure-function relationships (especially in disorder states).

A long term goal I have is to improve learning and memory in both humans and machines, in the process understanding how we can better treat learning and neurodevelopmental disorders, as well as atypical neurodegeneration. Outside of my research, I have an interest in developing DS/AI solutions unique to SG and applying temporal graphs to other domains (e.g. quantitative finance).

Recent Updates

  • Sep 2024: Our paper on evaluating the robustness of XAI algorithms for biomarker discovery from functional connectomes was accepted by IEEE BHI
  • May 2024: Uploaded a preprint on a review of fMRI biomarkers of neurological disorders discovered via GNN
  • May 2024: Our paper on learning modularity-guided structure-function interactions was accepted by MICCAI
  • Apr 2024: Presented our papers @ ICASSP 2024, held in Seoul (oral)
  • Apr 2024: Our paper on a 2-stage approach to ICH segmentation was accepted by the journal IEEE Access
  • Jan 2024: 2 abstracts accepted @ ISMRM 2024
    • Ensemble of GNNs for decoding dFC in ADHD (poster)
    • Interpretable GNN for classification of Parkinson’s disease using multimodal connectomes (oral)
  • Jan 2024: 2 abstracts accepted @ OHBM 2024
    • Analysing dFC biomarkers of ADHD subtypes (poster)
    • Structure-function interactions associated with fluid cognition via GNNExplainer (poster)
  • Dec 2023: 2 papers accepted @ ICASSP 2024
    • Generating subtype-specific biomarkers of Alzheimer’s disease via SplitGNN (oral)
    • Multimodal GNN for predicting fluid cognition (oral)
  • Nov 2023: Our paper on discovering site-specific salient features and site-invariant biomarkers of Schizophrenia was accepted by the journal Scientific Reports
  • Aug 2023: Our paper on subtype-specific biomarkers for autism was accepted by IEEE BHI (poster)
  • Aug 2023: Our paper on a network-based approach for modelling sMRI and PET scans was accepted by the journal Computers in Biology and Medicine
  • Jul 2023: Our paper on modelling multimodal connectomes using brain graphs and population graphs (BrainGAT), was accepted by IEEE BHI (rapid-fire presentation)