KeyNote

Anqi Qiu, National University of Singapore (NUS), Singapore

Short bio: Dr. Qiu is Deputy Head for Research & Enterprises at Department of Biomedical Engineering and director for BME Center at NUS Suzhou Research Institute. She is also affiliated at the Institute of Data Science, the N.1 Institute for Health at the National University of Singapore, and the Johns Hopkins University. She is a Master of Eusoff Hall, board member of NUS penal for Student Discipline at the National University of Singapore.

Topic: Laplace Operators and Graph Convolutional Neural Networks for Medical Applications

Abstract: The Laplace-Beltrami operator is a generalization of the Euclidean representation of the Laplace operator to an arbitrary Riemannian manifold. It is a self-adjoint operator and its eigenfunctions form a complete set of real-valued orthonormal basis functions. In this talk, I will introduce spectral Laplace-Beltrami wavelets and its computational algorithm. I will then demonstrate its use for smoothing and classification of the data defined on smooth surfaces embedded in the 3-D Euclidean space. Furthermore, I will discuss that the spectral Laplace-Beltrami Wavelets and Hodge-Laplacian to incorporate graph network topology can be used for the construction of geometric convolutional neural network (CNN). I will show the use of this method on brain morphology and functional networks for the prediction of Alzheimer’s Disease and Cognition in adolescents.

Brian Lim, National University of Singapore (NUS), Singapore

Short bio: Dr. Lim is an Assistant Professor of Computer Science at the National University of Singapore, and Principal Investigator at the Institute for Health Innovation & Technology (iHealthTech) and NUS Centre for Research in Privacy Technologies (N-CRiPT). He leads the NUS Ubicomp Lab that focuses on combining Machine Learning and Human-Computer Interaction to improve health, wellness and livability in smart cities with human-centric explainable AI. Brian received his PhD and MS in Human-Computer Interaction from Carnegie Mellon University, and a BS in Engineering Physics from Cornell University. He is the recipient of the Google Research Scholar Award (2022), CHI 2022 Best Paper Award, IMWUT Distinguished Paper Award 2019, and CHI 2009 Best Long Paper Nomination. He serves as Associate Chair for CHI (2018-Present) and Associate Editor for IMWUT (2019-2020).

Topic: Towards Human-Centered Explainable AI

Abstract: As AI pervades society, it is increasingly important to ensure their responsible use with eXplainable AI (XAI). While myriad explanation generation techniques have been proposed, they remain insufficient to support human understanding due to their poor fit and usability. I present several methods to make XAI more human-like by following processes from human philosophy and psychology, and demonstrate applications in medicine, privacy, and emotion. Furthermore, the real-world use of AI brings many requirements which can conflict with explainability. I discuss several confounds that diminish the effectiveness of XAI, such as cognitive load and privacy, and present methods to mitigate them. Therefore, by studying human cognitive processes and requirements, human-centric XAI can be made more interpretable and trustworthy.

Fang Liu, University of Notre Dame, United States

Short bio: Dr. Fang Liu is Professor and Associate Chair in the department of Applied and Computational Mathematics and Statistics at the University of Notre Dame, Indiana, USA. Dr. Liu holds a PhD in Biostatistics from the University of Michigan, Ann Arbor, USA. She is also an affiliated faculty member at the Harper Cancer Research Institute, the Eck Institute of Global Health, the Technology Ethics Center, and the Lucy Family Institute for Data and Society at Notre Dame. Dr. Liu’s main research interests include data privacy and differential privacy, Bayesian Statistics, statistical machine learning, missing data analysis, and applications of statistics to medical, biological, and social sciences, generously supported by the US National Science Foundation, the National Institutes of Health, Gates Foundation, UNITAID, and Notre Dame internal grants. She is co-Editor-in-Chief for ACM Transactions on Probabilistic Machine Learning and Associate Editor for Transactions on Data Privacy. Dr. Liu is an elected fellow of the American Statistical Association.

Topic: Privacy-preserving Inference of Group Mean Difference in Zero-inflated Right Skewed Data with Partitioning and Censoring

Abstract: We examine privacy-preserving inferences of group mean differences in zero-inflated right-skewed (zirs) data. Zero inflation and right skewness are typical characteristics of ads clicks and purchases data collected from e-commerce and social media platforms, where the protection of user privacy and individual data is critical. We develop likelihood-based and model-free approaches to analyzing zirs data with formal privacy guarantees. We first apply partitioning and censoring (PAC) to ``regularize’' zirs data to get the PAC data for better inferential utility and more robust privacy considerations compared to analyzing the raw data directly. We conduct theoretical analysis to establish the MSE consistency of the privacy-preserving estimators in the proposed approaches based on the PAC data and examine the rate of convergence in the number of partitions and privacy loss parameters. We conduct extensive simulation studies to compare the inferential utility of the proposed approaches and apply the methods to obtain privacy-preserving inference for the group mean difference in a real digital ads click-through dataset. Based on the theoretical and empirical results, we make recommendations regarding the usage of these methods in practice.