In this session, you will learn to set up and manage your Gradebook to accurately record and calculate student grades in the new Ultra Course View of HuskyCT. We will cover specific settings and features such as the different types of calculations, using categories, and working with the different Gradebook views that are available to instructors. We will also cover how to send a grade column to Student Admin for midterm and final grade submission. 09/24/2024 2:30pm - 3:30pm Register - https://fins.uconn.edu/secure_inst/workshops/workshop_view.php?ser=3166 (https://fins.uconn.edu/secure_inst/workshops/workshop_view.php?ser=3166)

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: Model Sparsification on Emerging Applications and TechnologiesAbstract: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.

UConn Sexpert Peer Support Drop-In Hours are a new, free service offered by peer health educators, the UConn Sexperts, on the UConn Storrs campus! Peer Support Drop-In Hours are a great option for students who have questions about sex and sexual health, are looking for a non-judgmental, laid-back environment to discuss a sex related concern or issue, or are interested in improving their sexual health and personal well-being. Our UConn Sexperts are trained to provide education, support, and connection to resources on and off-campus on a wide variety of topics pertaining to sex, sexual health, and relationships. Mondays: 11:00am-4:00pm                                                                                    Tuesdays: 3:00pm-6:30pm                                                                              Wednesdays: 10:30am-6:30pm                                                                                  Thursdays: 3:30pm-6:30pm                                                                                        Fridays: 10:30am-5:00pm UConn Sexperts (and supervising staff) are designated confidential employees under UConn's Title IX Reporting Obligations. Peer support sessions are for educational and support purposes only. Peer support visits are not on-call or emergency services, and are not for individualized medical advice, nor are they counseling or therapy. For more information, visit www.studenthealth.uconn.edu/sexperts 

Abstract. The solutions to Stefan problem with Gibbs-Thomson law (i.e., with surface tension effect) are well known to exhibit singularities which, in particular, lead to jumps of the associated free boundary along the time variable. The correct times, directions and sizes of such jumps are only well understood under the assumption of radial symmetry, under which the free boundary is a sphere with varying radius. The characterization of such jumps in a general multidimensional setting has remained open until recently. In our recent work with M. Shkolnikov and Y. Guo, we derive a separate (hyperbolic) partial differential equation — referred to as the cascade equation — whose solutions describe the jumps of the solutions to the Stefan problem in the absence of any symmetry assumptions. It turns out that a solution to the cascade equation corresponds to a minimal element of the set of equilibria in a family of (first-order local) mean field games. In this talk, I will present and justify the cascade equation, will show its connection to the mean field games, and will prove the existence of a solution to the cascade equation.

Interpersonal Group for Undergraduate Students Undergraduate Students Interpersonal Groups focuses on promoting emotional wellbeing as you balance academics, relationships, family, and personal responsibilities. Groups offer a supportive confidential space to share your concerns, practice skills and get feedback.To join this group therapy session, please call SHaW at 860-486-4705 (tel:860-486-4705). This session is held by Maritza Lugo-Stalker,  (https://studenthealth.uconn.edu/person/maritza-lugo-stalker/)   For many concerns that students face – like overwhelming stress, anxiety, difficult relationships, depression, academic difficulties, and more – group therapy is the best option for support and healing. Facilitated by Student Health and Wellness (SHaW) counselors, our therapy groups encourage peer support, promote emotional wellbeing, and increase a felt sense of connection. Participants often find that they feel less alone in their struggles, and walk away with newfound support and ideas for coping.