About

Sandy F. Chang is an Assistant Professor in the Department of History at the University of Florida. Her research focuses on Southeast Asia, Modern China, Colonialism, Migration, and Gender and Sexuality Studies. Her academic work involves exploring historical and contemporary issues related to these regions and themes, contributing to a deeper understanding of their social, political, and cultural dynamics.

Research topics

• Computer Science
• Artificial Intelligence
• Physics
• Thermodynamics
• Nuclear physics
• Telecommunications
• Chemistry
• Environmental science
• Materials science
• Systems engineering
• Mechanics
• Mathematics
• Data science
• Engineering
• Human–computer interaction
• Acoustics
• Statistics

Selected publications

• Predicting Thermomechanical Degradation in Bonded Interfaces Using Enhanced Image Processing and Deep Learning Techniques

  2025-05-27

  article1st authorCorresponding

  Detecting internal defects in power electronics packages is critical for its performance and reliability, especially under extreme operating conditions, as these defects can lead to catastrophic failure if not properly addressed. Confocal scanning acoustic microscopy (C-SAM) plays a key role in the nondestructive evaluation of bond layer degradation within a power electronics package by detecting defects such as delamination, voids, and cracks. However, accurately quantifying and predicting these defects from C-SAM images remains a significant challenge due to the high noise-to-signal ratio, which typically arises from both imaging process and bond patterns itself. In this paper, we explore machine learning strategies for processing C-SAM images and providing predictive models of defect growth. We use C-SAM images of sintered copper and sintered silver samples, which were obtained under accelerated thermal experiments, as the representative dataset for our study. We investigate the effect of Fourier transforms and wavelet transforms on these datasets to remove high-frequency noise and addressed noise across multiple scales with histogram equalization to enhance the contrast and improve the visibility of defects. As a result, defect boundaries can be clearly distinguished, enabling more accurate tracking of their growth over time. We then employ different time-series forecasting algorithms on the denoised images to formulate an image-based lifetime prediction model. Statistical models and deep learning techniques are trained on images obtained in the early stages of thermal shock, and defect growth in the later stages is predicted. Our work serves as a preliminary attempt to improve the accuracy of lifetime prediction models of power electronics packages, which is critical under extreme operating environments.

  PublisherDOI

• Machine Learning Boiling Prediction: From Autonomous Vision of Flow Visualization Data to Performance Parameter Theoretical Modeling

  SSRN Electronic Journal · 2024-01-01

  preprintOpen access
  PublisherDOI

• Modeling Flow Boiling Utilizing Machine Learning Vision Data

  2024-05-28

  article

  Flow boiling is a very efficient configuration for meeting the high heat dissipation demands of thermal management systems. However, the lack of a clear understanding of physics affecting two-phase flow has limited its wide implementation across thermal systems. Recently, novel machine learning vision tools have been developed to capture physical feature information during subcooled flow boiling in a rectangular channel with single-sided heating. These features include local and averaged, as well as transient and steady-state statistical data on void fraction, vapor-liquid voids, interfacial behaviors, and liquidsolid wall wetting front areas. In this study, this data is used to model performance parameters in flow boiling. The statistical information relating to void fraction, bubble behaviors, interfacial waviness, and wetting fronts is analyzed and correlated with heat transfer coefficients, and critical heat flux. The data is used in combination with traditional control-volume-based theoretical modeling techniques to capture the relationship between the parameter of interest and the input parameters. The data on channel void fraction and wetting front areas are used to improve theoretical models predicting heat transfer coefficient. The data on interfacial behaviors and wetting front areas is to improve mechanistic model predicting critical heat flux. With this work, a new approach to utilizing machine vision data is proposed and validated.

  PublisherDOI

• Constraining MeV to 10 GeV majoron by Big Bang Nucleosynthesis

  arXiv (Cornell University) · 2024-01-01

  preprintOpen access1st authorCorresponding

  We estimate the Big Bang nucleosynthesis (BBN) constraint on the majoron in the mass range between $1\,{\rm MeV}$ to $10\,{\rm GeV}$ which dominantly decays into the standard model neutrinos. When the majoron lifetime is shorter than $1\,{\rm sec}$, the injected neutrinos mainly heat up background plasma, which alters the relation between photon temperature and background neutrino temperature. For a lifetime longer than $1\,{\rm sec}$, most of the injected neutrinos directly contribute to the protons-to-neutrons conversion. In both cases, deuterium and helium abundances are enhanced, while the constraint from the deuterium is stronger than that from the helium. $^7{\rm Li}$ abundance gets decreased as a consequence of additional neutrons, but the parameter range that fits the observed $^7{\rm Li}$ abundance is excluded by the deuterium constraint. We also estimate other cosmological constraints and compare them with the BBN bound.

  PublisherOA PDFDOI

• Machine learning boiling prediction: From autonomous vision of flow visualization data to performance parameter theoretical modeling

  International Journal of Multiphase Flow · 2024-07-20 · 14 citations

  articleOpen access

  Flow boiling is a highly efficient configuration for meeting the high heat dissipation demands of thermal management systems. However, the complex physics of two-phase flow has hindered its broader application, especially in terms of quantifying visual information. Recent advancements in machine learning vision tools have revolutionized the analysis of phase change phenomena by enabling the digitalization of physically meaningful features such as void fraction, vapor-liquid interfacial behaviors, and liquid-solid wall wetting front areas en masse. In this study, we systematically investigate two-phase models that compute void fractions, heat transfer coefficients, and critical heat flux using live bubble data streams under microgravity. The collected empirical bubble data is used to supplement and validate traditional control-volume-based theoretical modeling approaches. Void fraction data is first validated with analytical frameworks. This is followed by void fractions and wetting front areas being used to improve correlations predicting heat transfer coefficients. This work showcases the potential of using a new machine learning-based strategy to accelerate scientific formula discovery through the extraction of multi-level and physically meaningful features.

  PublisherDOI

• Constraining MeV to 10 GeV Majorons by big bang nucleosynthesis

  Physical review. D/Physical review. D. · 2024-07-17 · 6 citations

  articleOpen access1st authorCorresponding

  We estimate the big bang nucleosynthesis (BBN) constraint on the majoron in the mass range between 1 MeV to 10 GeV which dominantly decays into the standard model neutrinos. When the Majoron lifetime is shorter than 1 sec, the injected neutrinos mainly heat up background plasma, which alters the relation between photon temperature and background neutrino temperature. For a lifetime longer than 1 sec, most of the injected neutrinos directly contribute to the protons-to-neutrons conversion. In both cases, deuterium and helium abundances are enhanced, while the constraint from the deuterium is stronger than that from the helium. <a:math xmlns:a="http://www.w3.org/1998/Math/MathML" display="inline"><a:mrow><a:mrow><a:mmultiscripts><a:mrow><a:mi>Li</a:mi></a:mrow><a:mprescripts/><a:none/><a:mrow><a:mn>7</a:mn></a:mrow></a:mmultiscripts></a:mrow></a:mrow></a:math> abundance gets decreased as a consequence of additional neutrons, but the parameter range that fits the observed <c:math xmlns:c="http://www.w3.org/1998/Math/MathML" display="inline"><c:mrow><c:mrow><c:mmultiscripts><c:mrow><c:mi>Li</c:mi></c:mrow><c:mprescripts/><c:none/><c:mrow><c:mn>7</c:mn></c:mrow></c:mmultiscripts></c:mrow></c:mrow></c:math> abundance is excluded by the deuterium constraint. We also estimate other cosmological constraints and compare them with the BBN bound. Published by the American Physical Society 2024

  PublisherDOI

• VISION-iT: A Framework for Digitizing Bubbles and Droplets

  Energy and AI · 2023 · 19 citations

  • Computer Science
  • Artificial Intelligence
  • Computer Science

  Quantifying the nucleation processes involved in liquid-vapor phase-change phenomena, while dauntingly challenging, is central in designing energy conversion and thermal management systems. Recent technological advances in the deep learning and computer vision field offer the potential for quantifying such complex two-phase nucleation processes at unprecedented levels. By leveraging these new technologies, a multiple object tracking framework called “Vision Inspired Online Nuclei Tracker (VISION-iT)” has been proposed to extract large-scale, physical features residing within boiling and condensation videos. However, extracting high-quality features which can be integrated with domain knowledge requires detailed discussions that may be field- or case-specific problems. In this regard, we present a demonstration and discussion of the detailed construction, algorithms, and optimization of individual modules to enable adaptation of the framework to custom datasets. The concepts and procedures outlined in this study are transferable and can benefit broader audiences dealing with similar problems.

  PublisherDOI

• BubbleMask: Autonomous visualization of digital flow bubbles for predicting critical heat flux

  International Journal of Heat and Mass Transfer · 2023 · 18 citations

  1st authorCorresponding
  • Computer Science
  • Artificial Intelligence
  • Mechanics
  PublisherDOI

• Autonomous Visualization of Digital Flow Bubbles for Predicting Critical Heat Flux

  SSRN Electronic Journal · 2023-01-01

  preprintOpen access1st authorCorresponding
  PublisherDOI

• VISION-iT: Deep Nuclei Tracking Framework for Digitalizing Bubbles and Droplets

  SSRN Electronic Journal · 2023-01-01 · 5 citations

  articleOpen access
  PublisherDOI

Frequent coauthors

• E. Conte

  Institut Pluridisciplinaire Hubert Curien

  218 shared

• D. Blöch

  Institut Pluridisciplinaire Hubert Curien

  218 shared

• J. Andreä

  Institut Pluridisciplinaire Hubert Curien

  216 shared

• C. Collard

  Institut Pluridisciplinaire Hubert Curien

  212 shared

• E. C. Chabert

  Institut Pluridisciplinaire Hubert Curien

  182 shared

• M. Lethuillier

  Institute of Nuclear Physics of Lyon

  181 shared

• P. Verdier

  Institute of High Energy Physics

  181 shared

• S. Perriès

  Institute of Nuclear Physics of Lyon

  172 shared

Awards & honors

• 2021 Berkshire Conference Article Prize in the history of wo…
• 2022 Nupur Chaudhuri First Article Award from the Coordinati…