Projects
Guided Hypothesis Exploration of Diseases Using Graph Mining
PI: Jaideep Vaidya
In the standard bioinformatics sense, researchers come up with a hypothesis based on doctor’s domain knowledge, and then carry out hypothesis testing to check if the hypothesis might hold or not. Our aim in this project is to, instead of relying on doctor’s domain knowledge, find interesting hypothesis of one specific general disease associated with other diseases or treatments. Then we check if they are meaningful or not through statistical analysis. Specifically, we will verify the predictive validity of graph mining, or to retest p-value/AUC after hypothesis tests.
Comparative Effectiveness of Intravenous Immunoglobulin for Treatment of Neurological Disorders
PI: Nizar Souayah
Intravenous immunoglobulin (IVIg) has been shown to be efficacious as second-line therapy in patients with inflammatory myopathies, myasthenia gravis, stiff person syndrome and multiple sclerosis. It was also effective in other autoimmune and inflammatory neuropathy. However, despite its effectiveness, the high cost of IVIG is a limiting factor for more extensive use of this drug. Our aim in this project is to compare the effectiveness of IVIG as compared to other therapeutic strategies in terms of length of hospitalization, disability, death.
Composition of Distributed Workflow Applications in the Cloud and Edge Computing Environment
PI: Basit Shafiq
The project aims at developing a framework for rapid development and deployment of distributed workflow applications using the resources available in the emerging cloud and edge computing environment. These resources include computational, data, and storage resources available in the cloud data centers and enterprise networks as well as large number of Internet of Things (IoT) devices providing diverse sensory and computation services. Development of distributed applications in such Internet-centric ubiquitous environment requires dynamic discovery, selection and integration of relevant services which could be available on the cloud data centers or deployed on IoT devices in the edge network.
Identifying Privacy Risks in Social Media
PI: Soon Ae Chun
Information sharing can be instantly achieved via social posts, chat messages, emails, blogs, etc., posing privacy as a major challenge in the hyper connected society. The disclosure of sensitive (or private) information about the individual user or others can occur intentionally or inadvertantly. In this project we present an approach to assess and analyze the privacy disclosure risk in the social media posts by a user when sharing information online. Our approach is to apply advanced ML and Deep Learning approaches to develop learning model to identify and detect in real-time whether a piece of information about to be shared contains sensitive information or not.
News Classification
PI: Soon Ae Chun
Nowadays on the Internet, there are a lot of sources that generate immense amounts of daily news. It is crucial that the news is classified to allow users to access effectively the information of interest. We apply Machine Learning, AI and Deep Learning Models to classify news into different categories.
Smart City: Economic Development Policy Analytics
PI: Soon Ae Chun
To design a good economic policies in a city, the visibility of the economic partners and their transactions and the prediction of economic growth potentials are essential. We develop a data analytics platform to analyze and predict the economic transactions among diverse agents in the city and allow governments to analyze their policy impacts.
- Integration of local supply and demand between anchor institutions and manufacturers/vendors
- Promote local manufacturers and vendors
- Increase employment opportunities
- Reduce crime rates
- Provide economic development analysis and forecast for the city mayor