Research Projects
Building Better Robot Brains
The main objective of this work is to improve machine intelligence which has applications in robotic survival, automated creative design, and management of complex systems like farms.
This is done through the application of computational cognition which is the study of how human brains work to inform the creation of autonomous systems.
Activities & Outcomes
- Improved Machine Intelligence
- Applications in
- Space Exploration
- Mission Command and Control
- Art Generation/Product Design
- Agriculture
- Funded By: NASA Nebraska EPSCoR and NSRI
Key Faculty and Collaborators
- Ada-Rhodes Short, Ph.D.
- Daniel Hulse, Ph.D., NASA Ames
- Dhundy Bastola, Ph.D., UNO
- Carl Nelson, Ph.D., UNL
Robotics, Networking and Artificial Intelligence (R.N.A.) Lab
RNA LAB Encompasses three areas: Cyber-Physical Systems, Networking and Computer Vision. The team vision is to explore innovative computing, networking, and machine learning/deep learning paradigms.
The lab is working to develop AI and related intelligent software robots capable of completing mission-critical tasks in unstructured environments without dependence on human guidance. They are creating intelligent networked systems, enabling resilient and secure communication, and are also studying how to improve machine vision to beyond human-level performance.
Activities & Outcomes
- Enhanced Imaging Frameworks Against Future Outbreaks of COVID-19
- A Global Soil Spectral Calibration and Prediction Network to Characterize the Global Soils
- A Real-Time Self-Adaptive Systems for Smart Building using Deep Reinforcement Learning and Wireless Sensor Networks
Key Faculty & Collaborators
- Pei-Chi Huang, Ph.D.
- Xin Zhong, Ph.D.
Knowledge Representation and Reasoning
Knowledge Representation and Reasoning (KR) is a subfield of Artificial Intelligence (AI) where a fundamental assumption is that an agent's knowledge is explicitly represented in a declarative form, suitable for processing by dedicated reasoning engines.
This assumption, that much of what an agent deals with is knowledge-based, is common in many modern intelligent systems. Consequently, KR has contributed to the the theory and practice of various areas in AI, including automated planning and natural language understanding, and to fields beyond AI, including databases, verification, software engineering, and robotics.
In recent years, KR has contributed also to new and emerging fields, including the semantic web, computational biology, cyber security, and the development of software agents.
Activities & Outcomes
- Received $300,000 funding from NSF
- More than 100 publications
- Autoepistemic answer set programming, Gelfond-Zhang
aggregates as propositional formulas, Forgetting auxiliary
atoms in forks, On abstract modular inference systems and
solvers, Relating constraint answer set programming
languages and algorithms, Constraint answer set
programming versus satisfiability modulo theories,
Information extraction tool Text2ALM: from narratives to
action language system descriptions
- Autoepistemic answer set programming, Gelfond-Zhang
Key Faculty & Collaborators
- Yuliya Lierler, Ph.D.
- Jorge Fandinno, Ph.D.
Deep Learning-based Foreign Object Debris Detection
Foreign Object Debris is any substance alien to an airport system that can cause damage. We leverage multidisciplinary techniques by integrating machine learning, computer vision, and small unmanned aerial technology to develop a system that helps overcome the high-cost, low-efficiency, and technical challenges of Foreign Object Debris detection that a great number of small-scale airports are facing.
Activities & Outcomes
- Received NU Collaboration Initiative and NASA Nebraska Space Grants
- One MS and two UG students supported
- Two papers published
- Developed one publicly accessible image dataset
- Developing external grant proposals collaborating with the Aviation Institute and the Geography/Geology Department at UNO
Key Faculty & Collaborators
- Xin Zhong, Ph.D.
Deep Weakly Supervised Learning
Labeling an image as a whole is much easier than labeling each object/segment inside the image. However, supervised learning can only learn from training data with exact labels. Hence, weakly supervised learning, more concretely, inexact supervision where the training data are given with labels but not as exact as desired, is the purpose of this project.
Here, the given label will be the classification labels, and the desired output will range from object localization to detection to segmentation.
Activities & Outcomes
- Received Nebraska University Research Development Program Grant
- Supported one Ph.D. student and one master's student
- Submitted two papers
Key Faculty & Collaborators
- Xin Zhong, Ph.D.
Deep Learning-based Image Watermarking
Image watermarking refers to covertly embedding information (i.e., a watermark) into a cover image. By designing deep learning methods such as specialized training schemes and novel layers, this project aims at adaptivity and robustness for image watermarking.
One typical downstream application is to extract a watermark from camera-resampled marked images, and the end users can scan any cover image for more information.
Activities & Outcomes
- Received NSF CRII Award (2104267)
- Supported one Ph.D. student, one master's student, and one undergraduate student worker
- Published four cited papers
- New Machine Learning course development at UNO
Key Faculty & Collaborators
- Xin Zhong, Ph.D.
Innovation in Targeted Violence and Terrorism Prevention
This project looked to understand issues around tips reporting at both local and national levels using a mixed methods approach. Multiple nationally representative surveys were conducted to identify underlying issues with reporting and evaluate multiple mechanisms to address those issues.
Results informed the development of an intelligent chatbot system which was iteratively evaluated and refined. Most immediately, the chatbot created from the proposed project would serve the greater Sarpy County, Nebraska population and the four University of Nebraska campuses located across the state (Omaha, Lincoln, Kearney, and the Medical Center).
Activities & Outcomes
- Received $715,000 from Center for Prevention Programs, and Partnerships (DHS)
- Community Engagement with Sarpy County Sheriff's Office
- Four graduate students and two undergraduate students supported this effort
Key Faculty & Collaborators
- Joel S. Elson, Ph.D.
- Erin Kearns, Ph.D.
Toward a Scalable Geo-distributed Data Analytics System
Analyzing highly diffused large-scale data, i.e., geodistributed data analytics (GDA), in a timely and cost($)-efficient manner, is one of the most important workloads for many internet applications. This project aims to design and implement a new and novel scalable GDA system that allows GDA applications to achieve desired cost-performance goals, which will allow them to provide better experiences to their users, e.g., user-centric content with reduced user-perceived latency.
Activities & Outcomes
- Received NSF CRII Award (2153422)
- Received Nebraska EPSCoR’s First Award
- Published a top-tier conference paper (ACM/IFIP Middleware’23)
- Supported one Ph.D. student and one master’s student
Key Faculty & Collaborators
- Kwangsung Oh, Ph.D.
Finding Software Reuse Errors
Our concrete goal is to (1) explore the abundance of predictable repetitive regularities of a massive codebase, (2) develop a machine learning approach for training a model to identify common patterns in software corpora, and then (3) use these patterns to detect anomalous, likely buggy, program behavior that significantly deviates from typical patterns.
Our approach captures internal regularities and repetitive properties of software as patterns and detects violations of common patterns.
Activities & Outcomes
- FY2024 Nebraska Research Initiative (NRI) Funds
- Supported one graduate assistant student
Key Faculty & Collaborators
- Myongkyu Song, Ph.D.
- Harvey Siy, Ph.D.
Maintaining Complex Changes to Support Quality Assurance
Code review and regression testing are important practices in software maintenance. However, it is not easy to use existing tools to inspect related changes that interfere with other changes. To address these problems, we have presented a novel approach to automation of change decomposition. Our approach automatically (1) decomposes changes of interest using data and control dependence relationships, (2) summarizes related changes by matching decomposed changes against the rest of a program, and (3) automatically applies identified related changes to the original program version to produce an intermediate source program version guaranteed to compile and run with test cases.
Activities & Outcomes
- Publication – COMPSAC '20, EIT '20, JCST '19, JIET '19, COMPSAC '19, SANER '19, JSEP ‘18, JSCP'23
- Student Support - 1 Ph.D. and 2 MS theses, GRACA 16', '19, FUSE 17', '19, '20
- Nebraska Research Initiative '15 with $270K
- Software tool demo at CS workshop '17 '19, program inspection and testing techniques for reliably evolving software
Key Faculty & Collaborators
- Myongkyu Song, Ph.D.
Network Congestion Control for a Large-Scale DSS
The disaggregated storage system (DSS) is a promising direction for a storage framework in a large-scale data center. It provides more flexible resource management, straightforward upgrade and maintenance, and other features that enterprise storage systems desire. The connecting network becomes critical to the entire system’s performance in a DSS.
Like traditional data center network, DSS requires extremely low latency to handle the storage I/O requests. Thus, network congestion becomes a significant concern in such a system. We aim to design solutions for network congestion issues in a DSS.
Activities & Outcomes
- Publications (Funded by Samsung):
- – ICNC ’23, IPDPS ‘23, IPCCC ’21
- Current Efforts:
- Design Reinforcement Learning-based network congestion control schemes for a DSS
Key Faculty & Collaborators
- Xiaoqian (Tiffany) Zhang, Ph.D.
Smart Wearable In-home Monitoring System for Health Aging
Falls in the US lead to 800,000 hospitalizations and 27,000 deaths yearly. The financial impact on older fallers may reach $101 billion by 2030 as the population ages. Rural older adults have been shown to be at greater risk of falls than non-rural residents, because they are often more isolated and receive less assistance after falling; lack of appropriate health care access may amplify falls among rural versus urban residents. Therefore, there is a critical need to develop a system using wearable devices to monitor mobility and physical health in rural elderly.
Activities & Outcomes
- Develop and evaluate a comprehensive set of mobility evaluation metrics for continuous mobility monitoring using SWIMS in older adults
- Refine and validate our machine-learning-based analysis model to investigate the relationships between clinical assessments and mobility evaluation metrics
Key Faculty & Collaborators
- Jon Youn, Ph.D. (UNO)
- Ka-Chun Siu, Ph.D. (UNMC)
- Jan Allison Moore, Ph.D. (UNK)
Fairness and Bias in Data-Driven Decision-Making
Automated, data-driven decisions are increasingly used in people’s everyday lives, from credit applications to healthcare diagnoses and even judicial decisions. Unfortunately, these approaches can perpetuate existing biases captured in the considered data and discriminate against underrepresented or disadvantaged groups.
This research project aims to define, identify, measure, and mitigate potential biases in various data-driven decision scenarios, ranging from prediction models for business or healthcare decisions to recommender systems and matching algorithms. It specifically focuses on what effect improving fairness has on other relevant decision criteria.
Activities & Outcomes
- International collaboration with colleagues from Ireland, Germany, and the US
- Several publications, including ACM CSUR, ICIS 2019, and Information Processing & Management
- Inclusion of Fairness and related challenges in ML courses
- Several Ph.D. and master theses
Key Faculty & Collaborators
- Christian Haas, Ph.D.
Computational Redistricting
States (generally, geographic areas) need to be divided into non-overlapping, contiguous voting districts to ensure that each vote is equally important (“One Person, One Vote”). Currently, setting up these districts is a topic of intense political debates. In many cases, partisan committees draw „unfair“, or „gerrymandered“ maps that give one party an over-proportional share of political representation.
This project aims to use algorithmic approaches to draw a population of different, legally valid maps that provide fair representation and that can be applied for federal, state, and local voting districts.
Activities & Outcomes
- Interdisciplinary collaboration with colleagues from Operations Research and Political Science
- Several publications
- Open-source toolkit and algorithms for computational redistricting that are scalable for federal, state, and local voting districts
Key Faculty & Collaborators
- Christian Haas, Ph.D.
- Steven O. Kimbrough, Ph.D. (Wharton School, UPenn)
- Peter Miller, Ph.D. (NYU Brennan Center for Justice)
Multiple Objective Optimization in Computer and
Wireless Sensor Networks using Computational Intelligence
In this project we seek to study computer and wireless networks optimization using computational intelligence, evolutionary computation and swarm intelligence. Past research includes the optimization of multicast networks and wireless sensor networks placement to find places to minimize communication/energy usage and deployment costs.
Activities & Outcomes
- Papers published in conferences such as IEEE
- Symposium in Multicriteria Decision Making, IEEE
- WCNC and IEEE AINA
Key Faculty & Collaborators
- Alfredo J. Perez, Ph.D. (UNO CS Lead)
Innovation for Domestic Terrorism Prevention: Investigating the Use of Chatbots in Suspicious Activity Reporting
This project identifies challenges and problems associated with suspicious activity reporting, develop and test innovative technologies to help alleviate these tensions and pain-points.
Partnered with Sarpy County Sheriff’s Office & Awareity. Conducted focus groups with threat assessment teams & analyzed reporting data, trends, and behavior. Developed and experimentally tested efficacy of chatbots in comparison to traditional reporting options.
Activities & Outcomes
- Funded by DHS Center for Prevention Programs and Partnerships under Grant Award Number DHS 21-TTP-132-00-01.
- Academic & practitioner publications
- Awarded US Patent
- Selected for federal technology accelerator initiative
- Supported 8 Graduate Assistants and 5 Undergraduates
Key Faculty & Collaborators
- Joel S. Elson, Ph.d.
- Erin M. Kearns
AI in Mixed Initiative Teams: Creativity and Collaborative Outcomes
Project looked to understand the role of AI as malign tool in facilitating novel attacks and advance empirical work on what aspects of technology impact trust, collaboration, and malevolent creativity.
Activities & Outcomes
- Funding: Cooperative agreement (Grant Award Number 20STTPC00001-04). DHS Office of University Programs.
- Practitioner oriented scientifically grounded summary of AI as malign tool.
- Empirical studies on the use of AI for creativity in mixed initiative teams.
- Supported: 3 Graduate Assistants and 2 Undergraduates
Key Faculty & Collaborators
- Joel S. Elson, Ph.d.
- Sam T. Hunter