Proposal Description

This invited session presents a compelling array of cutting-edge statistical methodologies and theories, meticulously crafted to tackle the intricate challenges posed by complex data scenarios. Spanning functional data, distributions, spherical data, networks, and data residing in a general metric space, the session promises to offer profound insights into innovative statistical approaches that find applications across various disciplines, including economics, geophysics, energy, sociology, etc.

This proposed session showcases the expertise of four distinguished speakers, comprising a balanced mix of three female and one male researchers. Hailing from prestigious universities across continents, these speakers range from junior to senior researchers, each contributing unique perspectives and insights to the topics.

Speaker 1: Jane-Ling Wang, Distinguished Professor, University of California, Davis, USA
Title: Learning theory for functional data
Deep neural networks have emerged as powerful tools across various domains, yet their potential for functional data remains underexplored. Dr. Wang endeavors to bridge this gap by exploring the challenges of leveraging deep neural networks for nonparametric function regression. Her research aims to establish theoretical underpinnings for deep learning approaches in functional data analysis, promising improved convergence rates and adaptability to complex data structures.

Speaker 2: Juan Cuesta-Albertos, University Professor, Universidad de Cantabria, Spain
Title: Do individuals in this population have higher incomes than those in that population?
Dr. Cuesta-Albertos introduces novel indices to assess differences in income distributions between populations, moving beyond simplistic comparisons of global parameters such as population means. By utilizing stochastic order and percentiles, his approach offers nuanced insights into income disparities within and between populations.

Speaker 3: Janice Scealy, Associate Professor, Australian National University, Australia
Title:  Exploration of geophysical data using geometry driven statistical methods
Dr. Scealy navigates the statistical analysis of complex geophysical datasets, ranging from palaeomagnetism to seismology. By integrating hierarchical sample designs and manifold data models, she showcases the efficacy of geometry-driven statistical methods in tackling geophysical challenges. Her research provides valuable tools and insights for understanding the intricate patterns and dynamics inherent in geophysical data.

Speaker 4: Paromita Dubey, Assistant Professor, University of Southern California, USA
Title: Change point detection for random objects using distance profiles
Dr. Dubey introduces a powerful scan statistic for detecting change points within data sequences that reside in a separable metric space. Leveraging distance profiles, her non-parametric approach offers precise change point detection and estimation, with applications across diverse data types.

This session promises to illuminate the frontiers of statistical research, offering profound insights into advanced methodologies and their applications across diverse data landscapes. Attendees will gain invaluable knowledge tailored for addressing contemporary data challenges, with a broad exploration of various types of complex data.