A5068819570- Data-Driven Disease Surveillance
- Spatial and Panel Data Analysis
- Bayesian Methods and Mixture Models
- Privacy-Preserving Technologies in Data
- Bayesian Modeling and Causal Inference
- Stochastic Gradient Optimization Techniques
- Gaussian Processes and Bayesian Inference
- Economic and Environmental Valuation
- Healthcare Policy and Management
- Advanced Graph Neural Networks
- Data Visualization and Analytics
- Robotic Locomotion and Control
- Animal Behavior and Welfare Studies
- Topic Modeling
- Gait Recognition and Analysis
- Statistical Methods and Bayesian Inference
Queensland University of Technology
2023-2025
Murdoch University
2024
Cancer Council Queensland
2024
Background Spatial data are often aggregated by area to protect the confidentiality of individuals and aid calculation pertinent risks rates. However, analysis spatially is susceptible modifiable areal unit problem (MAUP), which arises when inference varies with boundary or aggregation changes. While impact MAUP has been examined previously, typically these studies have focused on well-populated areas. Understanding how behaves sparse particularly important for countries less populated...
Predictions of animal movement are vital for understanding and managing wild populations. However, the fine-scale, complex decision-making animals can pose challenges accurate prediction trajectories. Step selection functions (SSFs), a common tool inferring relationships between environment, also increasingly used to simulate trajectories prediction. Although admitting lot flexibility, SSF framework is limited its reliance on pre-defined functional forms fitting data, SSFs that involve model...
Building on a strong foundation of philosophy, theory, methods and computation over the past three decades, Bayesian approaches are now an integral part toolkit for most statisticians data scientists. Whether they dedicated Bayesians or opportunistic users, applied professionals can reap many benefits afforded by paradigm. In this paper, we touch six modern opportunities challenges in statistics: intelligent collection, new sources, federated analysis, inference implicit models, model...
Abstract Background Spatial data are often aggregated by area to protect the confidentiality of individuals and aid calculation pertinent risks rates. However, analysis spatially is susceptible modifiable areal unit problem (MAUP), which arises when inference varies with boundary or aggregation changes. While impact MAUP has been examined previously, typically these studies have focused on well-populated areas. Understanding how behaves sparse particularly important for countries less...
Federated learning (FL) allows for collaborative model training across decentralized clients while preserving privacy by avoiding data sharing. However, current FL methods assume conditional independence between client models, limiting the use of priors that capture dependence, such as Gaussian processes (GPs). We introduce Structured Independence via deep Generative Model Approximation (SIGMA) prior which enables non-factorizable models clients, expanding applicability to fields spatial...
Spatial statistical models are commonly used in geographical scenarios to ensure spatial variation is captured effectively. However, and cluster algorithms can be complicated expensive. One of these geographically weighted regression (GWR) which was proposed the geography literature allow relationships a model vary over space. In contrast traditional linear models, have constant coefficients space, estimated locally at spatially referenced data points with GWR. The motivation for adaption...
Vertical federated learning (VFL) has emerged as a paradigm for collaborative model estimation across multiple clients, each holding distinct set of covariates. This paper introduces the first comprehensive framework fitting Bayesian models in VFL setting. We propose novel approach that leverages data augmentation techniques to transform problems into form compatible with existing algorithms. present an innovative formulation specific scenarios where joint likelihood factorizes product...
Federated learning methods enable model training across distributed data sources without leaving their original locations and have gained increasing interest in various fields. However, existing approaches are limited, excluding many structured probabilistic models. We present a general elegant solution based on variational inference, widely used Bayesian machine learning, adapted for the federated setting. Additionally, we provide communication-efficient variant analogous to canonical...
This article provides a comprehensive synthesis of the recent developments in synthetic data generation via deep generative models, focusing on tabular datasets. We specifically outline importance context privacy-sensitive data. Additionally, we highlight advantages using models over other methods and provide detailed explanation underlying concepts, including unsupervised learning, neural networks, models. The paper covers challenges considerations involved for datasets, such as...
Spatial statistical models are commonly used in geographical scenarios to ensure spatial variation is captured effectively. However, and cluster algorithms can be complicated expensive. This paper pursues three main objectives. First, it introduces covariate effect clustering by integrating a Bayesian Geographically Weighted Regression (BGWR) with Gaussian mixture model the Dirichlet process model. Second, this examines situations which particular holds significant importance one region but...
Building on a strong foundation of philosophy, theory, methods and computation over the past three decades, Bayesian approaches are now an integral part toolkit for most statisticians data scientists. Whether they dedicated Bayesians or opportunistic users, applied professionals can reap many benefits afforded by paradigm. In this paper, we touch six modern opportunities challenges in statistics: intelligent collection, new sources, federated analysis, inference implicit models, model...