(1) Computational genomics
We answer the question of how can our reliability techniques be used to make the promise of genomics-based personalized medicine reliable. In this domain, we look at how to develop programing framework and runtime environment for executing demanding computational genomics applications reliably. How can we use machine learning algorithms to understand the patterns of errors made by genomic instruments and automatically compensate for such errors?
Read More
(2) Security in social networks
Can we apply streaming machine learning algorithms to determine security failures in our social networks? What can we do to suppress the noise inherent in such data streams? Our solution is looking at applying ML and NLP techniques to the problem while attaining goals of speed, interpretability, and low false positives.