More about HKUST
A Survey of Representation Learning for Volumetric Scalar Fields
PhD Qualifying Examination
Title: "A Survey of Representation Learning for Volumetric Scalar Fields"
by
Mr. Bolin ZHAO
Abstract:
Volumetric scalar fields and their time-varying counterparts are widely
produced across various scientific and engineering domains, such as physical
oceanography, climate science, and manufacturing. Effectively representing
this data is crucial for facilitating computation, storage, visualization, and
analysis. In recent years, advances in learning-based methods have introduced
a broad range of frameworks for representing scalar fields to support diverse
tasks, such as generation, compression, and interactive visualization. This
diversity of options necessitates a structured overview of their
methodological foundations, supported tasks, and application settings. This
survey provides a comprehensive and critical review of recent progress in
learning-based representations for scalar fields. Following a structured
literature collection and selection process, we organize the reviewed work
into four major methodological categories: implicit neural representations,
convolution-based methods, transformer-based methods, and Gaussian splatting.
For each category, we examine how the corresponding approaches represent
steady and unsteady scalar fields, the scientific tasks they support, and the
settings in which they are applied. We further synthesize their technical
characteristics, application objectives, and evaluation practices to present
the relationships among different representation paradigms. Through this
review, we aim to provide an accessible account of the development and current
landscape of learning-based volumetric data representation, support the
selection and comparison of representation approaches for different scientific
tasks, and outline opportunities for continued research in this area.
Date: Tuesday, 29 September 2026
Time: 9:00am - 11:00am
Venue: Room 5501
Lift 25/26
Committee Members: Dr. Xiaojuan Ma (Supervisor)
Dr. Jun Han (Co-Supervisor)
Dr. Sehi L'Yi (Chairperson)
Dr. Arpit Narechania