The explosive growth of biological data now far exceeds the pace at which computer technologies are evolving to manage and use this information. The Health Data Analytics Center was established by the College of Engineering at Illinois to address the critical needs for data-driven approaches to bring descriptive, predictive, and prescriptive power to scientists, physicians, patients, and caregivers.
Many bioinformatics algorithms rely on optimization and complex statistical computations and do not scale to the massive biological and clinical datasets that now exist. The Center brings together a multidisciplinary team of researchers drawn from the fields of medicine, biology, public health, informatics, computer science, applied mathematics, and statistics. Together, our mission is to improve the health of individuals and the healthcare system through data-driven methods and understanding of health processes—to engineer advanced descriptive, predictive, and prescriptive methods for medical disorders to the benefit of researchers, physicians, patients and caregivers.
The Center will address three major challenges:
1. Develop large and standardized medical data: Data-driven techniques including the cutting edge deep learning/AI techniques require large and well-curated datasets from across the health care landscape.
2. Advanced analytical techniques for healthcare: Expand data-driven and machine learning techniques curate and harmonize large datasets. Computer scientists scale the infrastructure and algorithms; visualization experts enhance user understanding of the data, and interdisciplinary teams of physicians, biologists, clinicians validate and interpret the algorithms and results. Ultimately, these techniques could lead to new robotics and automation systems
3. Significant computation and storage resources: Big data and AI demands huge datasets and computation to create appropriate models, often exceeding the capacity of conventional health care community computing facilities even though the inferencing engines applying the models can be mass produced inexpensively