Human Development & Family Sciences Ph.D.

The Intergenerational Transmission of Emotional Well-Being in Chinese Families and the Role of Filial Piety

‘Electrical Breakdown in Polymers: Insights from the Noninsulating Properties of Insulating Materials’

HDFS PhD Program

“The Role of the Internet on Sexual and Gender Diverse Youth’s Identity Development”

History Department faculty, graduate students, friends and family are invited to attend the dissertation defense of Abdullah Alhatem, who will be speaking and answering questions about his Ph.D. thesis: “The Domestic Slave Trade and the “Christian Master” in New England”

Studies of the electronic structure of 2D TMDs: Vanadium Dichalcogenides (VX2)
Condensed matter physics

Thesis Title: Precision Measurements of the Neutron Magnetic Form Factor to High Momentum Transfer using Durand’s Method
Field of Study: Physics

Differential Cross Sections from CLAS12 RG-A Inclusive Electron Scattering

Title: Novel Mechanisms of Chronic Kidney Diseases
Field of Study: Biomedical Science Ph.D. Program

Title: Model Sparsification on Emerging Applications and Technologies

Ph.D. Candidate: Shanglin Zhou

Major Advisor: Dr. Caiwen Ding
Co-Major Advisor: Dr. Krishna Pattipati
Associate Advisors: Dr. Cunxi Yu, Dr. Zhijie Shi

Date/Time: Tuesday, September 24th, 2024, 3:00pm - 4:00pm
Location: Virtual

Abstract:

Deep neural networks (DNNs) with higher accuracy often lead to larger models, increasing storage and energy demands. Model sparsification can reduce size but risks compromising accuracy. Balancing these factors is challenging, especially as traditional processors struggle with the requirements of low-power, real-time DNNs, highlighting the need for more efficient solutions.

In this thesis, we explore model sparsification in emerging applications. We propose an optimization approach using Surrogate Lagrangian Relaxation (SLR) for weight sparsification, streamlining the typical time-consuming three-step pipeline. It enables faster convergence and maintains high accuracy, even during hard-pruning, with rapid recovery in retraining.

We explore two primary emerging applications: (1) Energy-harvesting devices that require dynamic power management. We introduce EVE, an AutoML framework using SLR-based sparsification to find optimal multi-models with shared weights, reducing memory use and adapting to changing environments. (2) Diffractive Optical Neural Networks (DONNs) that are fast and energy-efficient but suffer from accuracy degradation due to interpixel interactions. We propose a physics-aware optimization framework for DONNs, incorporating SLR-based sparsification and roughness modeling to smooth phase changes and preserve accuracy.

Additionally, we extend our research to multi-task learning (MTL) with DONNs, which traditionally requires manual reconfiguration. We propose LUMEN-PRO, an automated MTL framework that utilizes a flexible DONN backbone. By rotating shared layers instead of storing task-specific ones, LUMEN-PRO reduces memory usage and enhances accuracy, enabling efficient and high-performance DONNs across various tasks.

Title: Catching Caribou in the Age of Climate Change: How Regulation and Caribou Population Decline Threaten the Alaska Native Way of Life
Field of Study: Anthropology

Title: Investigation of thyrotropin-releasing hormone neurons in the mouse lateral hypothalamic area
Department: Physiology and Neurobiology