Research Focus

CARIDA will leverage the research expertise and technical leadership of its members to carry out research across the “full stack” of research areas relevant to AI/ML/BD, ranging across (1) Big Data Storage and Management, (2) Artificial Intelligence and Machine Learning, and (3) Applications.

Big Data Storage and Management

Modern data generation technologies generate massive amounts of heterogeneous data streams that need to be stored, analyzed, interpreted and appropriately leveraged to facilitate decision-making and control. One example of such technologies is sensors and cyber-physical systems (e.g., IoT) which span many application domains including environmental, healthcare, smart automobiles, and smart cities. Traditional data storage and management systems are unable to cope with the data management needs of modern applications. As an example, we have been working on methods of optimizing data index structures to leverage main memory for minimizing access of slow disks and to leverage CPU cache for reducing access of slow DRAMs. High-performance computing systems that rely on fast networks and distributed and multi-core computers, as well as emerging cloud computing platforms, are also critical components in Big Data analytics solutions. In particular, our recent projects studied schemes of coordinating cross-server data placement and request scheduling to maximize aggregate data access throughput in computer clusters.

Artificial Intelligence and Machine Learning

While many data mining and machine learning algorithms have been around for a long time, the ability to use them in many big data applications is a recent development. In this regard, AI and machine learning are going through a revolution and the recent advances – e.g., new ML algorithms such as Deep Neural Networks, as well as advances in hardware, sophisticated algorithms and scalable software frameworks - are overturning traditional approaches to data analytics. In current projects, we train deep learning models on terabytes of video data to produce modules that perform human motion analysis, for applications such as sign language recognition and scoring of performance in physical exercises.

Applications

Recent advances in AI/ML/BD have revolutionized many areas of research and applications, such as computer vision, natural language processing, autonomous cars, robotics, genomics, and many areas of physical and social sciences, engineering, healthcare, and business. For instance, our inter-disciplinary research program in Computational Journalism has made UTA a leading institution in this field. In this area we are conducting projects on data-driven fact-checking, exceptional fact discovery, and health misinformation mitigation. For our Big Data Analytics applications, we expect to leverage various high performance computing facilities available to the Center, including the UT-wide TACC facilities, as well as UTA’s own HPC facilities.