The Cool Stuff
Under the impeccable guidance of Dr. Maurizio Filippone at EURECOM, my primary research interests involve developing scalable approximations to Gaussian process inference without compromising on performance and precision.
Of particular interest is whether the compromise on performance associated with such approximations can be tuned to a given computational budget.
The resilient appeal of Bayesian inference models in the face of competitive deep learning techniques can be attributed to their well-founded quantificaiton of uncertainty.
Bridging the gap between Gaussian processes and deep learning techniques remains a pertinent research goal.
My research has resulted in publications appearing in top-tier Machine Learning conferences including ICML and UAI.