Software engineering research projects

Research Experience for Undergraduates (REU): Data and software program application

Center for Integrative Neuroscience: Virtual Reality and Augmented Reality Core

  • NIH: COBRE
  • PI: Eelke Folmer
  • $655,380
  • September 2017 - May 2022

Optimizing Deep Learning Training through Modeling and Scheduling Support

  • NSF
  • PI: Feng Yan
  • $175,000
  • June 2018 - May 2020

Deep learning models trained on large amounts of data using lots of computing resources have recently achieved state-of-the-art training performance on important yet challenging artificial intelligence tasks. The success of deep learning has attracted significant research interest from hardware and software communities to improve training speed and efficiency. Despite the great efforts and rapid progress made, one important bridge to connect software and hardware support with deep learning domain knowledge is still missing: efficient configuration exploration and runtime scheduling. Both the quality of deep learning models and the training time are very sensitive to many adjustable parameters that are set before and during the training process, including the hyperparameter configurations (such as learning rate, momentum, number and size of hidden layers) and system configurations (such as thread parallelism, model parallelism, and data parallelism). Efficient exploration of hyperparameter configurations and judicious selection of system configurations is of great importance to find high-quality models with affordable time and cost. This is however a challenging problem due to a huge search space, expensive training runtime, sparsity of good configurations, and scarcity of time and resources.

The objective of this research work is to systematically study the unique properties of deep learning systems and workloads, and establish new modeling and scheduling methodologies for improving deep learning training. The PI aims to improve the efficiency of discovering high performing models through a dynamic scheduling methodology driven by a novel hyperparameter configuration classification approach. The PI aims at developing an accuracy- and efficiency-aware hybrid scheduling methodology that makes judicious scheduling decisions based on a global view of both the time dimension (accuracy potential) and spatial dimension (efficiency potential) information. This research work integrates techniques in workload characterization, performance modeling, resource management, and scheduling to dramatically speedup the training process while significantly reducing the cost in time and resources. More broadly, this project will gain foundational knowledge about the interaction between software-hardware support and deep learning domain knowledge. This knowledge can help design next generation deep learning systems and frameworks, making deep learning training handy for researchers and practitioners with limited system and machine learning domain expertise. This research will help enhance curriculum and provide research topics for both undergraduate and graduate students, especially students from underrepresented groups.

The Solar Energy-Water-Environment Nexus in Nevada

  • NSF EPSCoR
  • Co-PIs: Sergiu Dascalu, Fred Harris
  • $20,000,000 + $4,000,000 state match
  • August 2013 -  Nov 2018

In this five-year project, Nevada Experimental Program to Stimulate Competitive Research (NV-EPSCoR) addresses critical practical problems of relevance to large-scale solar installations in arid desert lands. The project combines research on solar thermal energy generation with the understanding of eco-hydrological impacts of solar installation in desert regions to advance the economic and eco-friendly viability of solar electricity generation. This combination distinguishes this project from several other existing solar energy projects, thus making it a unique model study of relevance to Nevada and other solar installations in the US and around the world. The major participating institutions in this project are: the University of Nevada, Reno, the University of Nevada, Las Vegas, and the University of Nevada Desert Research Institute. Faculty and students from the College of Southern Nevada, Truckee Meadows Community College, and Nevada State College will also be engaged in this project.

Intellectual merit

Despite plentiful sunlight and cloud-free days which are conducive for solar energy collection, arid regions experience frequent dust storms and receive little or no rain. Dust accumulated on solar panels absorb sunlight and decrease the efficiency of solar cells; water scarcity increases the cost of meeting the cooling needs of solar thermal collectors. This project seeks to develop engineering/technological solutions to repel dust and minimize water usage in large solar installations. In addition, it examines the desert ecosystem responses and provides science-based information for designing effective ways to manage and mitigate environmental impacts associated with large-scale solar installations. The award establishes a research facility, called the Nevada Environment, Water, and Solar Testing and Research Facility (NEW-STAR) during this project. Enhancements to the existing cyberinfrastructure capabilities are effected through the creation of the Nevada Research Data Center (NRDC) for data management and communication. These new facilities promote collaboration among teams of interdisciplinary scientists and engineers on solar-energy-water-environment nexus research and education theme.

Broader Impacts

This project has the potential to develop less costly and thus more competitive solar electricity generation techniques aimed at minimizing both water usage and environmental degradation. The technological solutions to be developed are applicable to other solar energy installations nationally and globally. Interactions among scientists and collaboration at regional, national, and international levels as well as partnerships with energy industry and environmental agencies in Nevada are expected to promote economic development in the state. The project involves 41 faculty, 24 technicians, 43 graduate students, and 38 undergraduate students as participants. The project offers the following programs aimed at pre-college and undergraduate students with a focus to attract underrepresented minority groups, K-12 teachers, and the general public:

  1. Pre-college bridging programs to assist K-12 students to develop academic skills and career pathways
  2. Undergraduate research opportunity programs (UROP) to provide research experience at the solar energy-water-environment nexus
  3. Undergraduate and graduate hands-on training (HOT) to facilitate transition from student to professional; activities include industry internships and laboratory experiences in solar energy technology, and proposal writing workshops
  4. Teacher professional development programs that engage K-12 teachers in research, field work, and working with graduate students
  5. Programs to educate K-12 students on project related themes and inform their families of opportunities for Science, Technology, Engineering, and Mathematics (STEM) careers
  6. Online learning laboratories, which provide wireless access to cyber learning materials and enhance public understanding of solar energy and related impacts on water and environment.

Understanding Gender Differences in Visual/Vestibular Conflict during Virtual Locomotion

  • Google Research
  • PI: Eelke Folmer Co-PI: Paul MacNeilage and Lars Strother
  • $44,044
  • July 2018 - June 2019