IT Faculty Research Seminar Series
Seminar # 6 - March 23rd, 2020. Recorded Presentation.
Presenter - Dr. Hossain Shahriar
Title: Security and Privacy in Open Source EHR Applications
Abstract: Electronic Health Record (EHR) applications are digital versions of paper-based patient's health information. They are improving the quality in healthcare delivery with easy access patient medication history, prior clinic visits, treatment plans, and precise decision-making process.
Before an EHR application can be used by clinics and hospitals, it need to be certified based on Office of the National Coordinator (ONC) guidelines which has overlapping data privacy and security requirements with Health Insurance Portability and Accountability Act (HIPAA). Unfortunately, there is no single tool available to automatically check the compliance of security and privacy requirements for a given EHR application.
In this talk, we address this problem. First, we provide an overview of Protected Health Information (PHI), HIPAA/ONC data security requirements. Then, we introduce two large scale open source EHR systems (OpenEMR and OpenClinic), and two static analysis tools (RIPS and VulnHunter). The static analysis tools were originally designed to detect security bugs based on OWASP guidelines. We then map the analysis results with HIPAA/ONC requirement checking. The initial results show that open source EHR systems suffer from security bugs leading to non-compliance of HIPAA and ONC requirements. Further, utilizing multiple analysis tools lead to better outcome towards compliance checking..
Seminar # 5 - Feb 27th, 2020 12pm-1pm at J-215B
Presenter 1 - Dr. Ming Yang
Title: Scale/Rotation/Perspective/Duplication Invariant Template Matching. Description: Template matching is a process that is widely applied in compute vision fields, such as location recognition, image panoramas, object recognition/tracking, image stitching, image retrieval, virtual environment navigation, etc. Traditional template matching algorithm is pixel based, which compares template and reference pixel by pixel, but it fails when rotation/scaling/distortion are present. Speeded Up Robust Reatures (SURF) has introduced feature-based template matching, which identifies feature points and utilizes feature point descriptor to find the match between template and reference. However, SURF does not always get good matching if there are similar patterns existing in the template/reference. In this study, we are trying to improve the performance of SURF algorithm, by introducing high-pass filtering to the images, as well as feature points filtering. We also proposed a feature point interpolation algorithm, to achieve template matching in scenarios where template and reference images are not temporally synchronized.
Presenter 2 - Dr. Jack Zheng
Title: Data Visualization and Education. Description: Dr. Zheng is developing two new streams of research and creative activities recently: 1) interactive data and information, with a focus on data visualization and interaction on emerging media such as video walls and VR/AR environments; 2) K-12 data technology literacy education and curriculum development, with a focus on incorporating the visual and art elements in technologies. Both of the streams are based on his years of experience and research in related fields. In this seminar, he will provide an overview of these two related streams, and some initiation projects.
Seminar #4 - Jan 24 12pm-1pm Location: J-215B
Presenter #1. Dr. Ying Xie
Title: Deep KNN and Its applications in Fintech and Healthcare
Description: The k Nearest Neighbor (KNN) algorithm has been widely applied in various supervised learning tasks due to its simplicity and effectiveness. However, the quality of KNN decision making is directly affected by the quality of the neighborhoods in the modeling space. Efforts have been made to map data to a better feature space either implicitly with kernel functions, or explicitly through learning linear or nonlinear transformations. However, all these methods use pre-determined distance or similarity functions, which may limit their learning capacity. In this talk, we present a novel deep learning method called Deep KNN that is able to learn pairwise similarities of data, which implicitly maps data to a feature space where the quality of KNN neighborhoods is optimized. We will further demonstrate successful applications of the proposed Deep KNN in critical domains such as Fintech and Healthcare with superior performance.
Presenter #2. Dr. Zhigang Li
Topic: Technology Enabled Learning
Description: Dr. Li will briefly introduce his research interests in technology enabled learning and then present a case study on a research-based course development framework that was implemented for an Intro to FinTech course.
Serminar #3 November 21, 4pm-5pm. Marietta Campus J-215B.
Dr. Seyedamin Pouriyeh
Abstract: In this talk, Dr. Pouriyeh will focus on two main research areas including exploration of RDF data using topic modeling technique and security issues in federated machine learning. Dr. Pouriyeh will present his recent conference and journal papers and his ongoing research topics and projects.
October 29th, 2019
4pm-5pm at J-132
Dr. Meng Han
Title: Security of Artificial Intelligence in Cloud Environment
Brief Description: An Empirical Study. Description: By utilizing deep learning techniques, most of the major cloud providers begin to offer services such as image auto-classification, object identification, and illegal image detection, etc. In this talk, Dr. Han is going to present an empirical study on black-box attacks to fool cloud-based image detectors. Dr. Han’s research shows that all major cloud platforms are vulnerable to the proposed blackbox attacks and possible defenses to such security challenges will be discussed.
Dr. Shirley Tian
Title: Two research streams: Behavioral Aspects of Cybersecurity and social media analytics.
Brief description: Dr. Tian will present two research streams in behavioral aspects of cybersecurity and social media analytics. Two detailed research in each stream will be introduced. And Dr. Tian will briefly introduce her ongoing research topics and projects.
September 23th, 2019
4pm-5pm at J -157
Dr. Richard Halstead-Nussloch
Title: Rich H-N’s Research
Brief Description: Over fifty years of successful sponsored research … (to be continued)
Dr. Lei Li
Title: Developing A Blockchain-Enabled Collaborative Intrusion Detection System
Abstract: Collaborative intrusion detection system (CIDS), where IDS hosts work with each other and share resources, have been proposed to cope with the increasingly sophisticated cyberattacks. Despite the promising benefits such as expanded signature databases and alert data from multiple sites, trust management and consensus building remain as challenges for a CIDS to work effectively. The blockchain technology with built-in immutability and consensus building capability provides a viable solution to the issues of CIDS. In this paper, we introduce an architecture for a blockchain-enabled signature-based collaborative IDS, discuss the implementation strategy of the proposed architecture and developed a prototype using open-source projects such Hyperledger and Snort. Our preliminary evaluation on a benchmark showed the proposed architecture offers a feasible solution by addressing the issues of trust, data sharing and insider attacks in the network environment of CIDSs. The implications and limitations of this study are also discussed.