Probabilistic Computing

Probabilistic Computing The recent increase in the volume and variety of data has raised questions about the methodology used to conduct real-life data analysis. Statistical inference problems include detecting predictive relationships between variables inferring missing values and identifying statistically similar database entries. BayesDB is a platform for data analysis that is built on a non-parametric Bayesian meta-model that automatically infers a series of mixture models from datasets. I will focus on the evaluation of BayesDB with regards to two of its chief functions: (1) its ability to detect dependencies between variables in datasets and (2) its ability to retrieve and order database records by similarity to known or hypothetical records of interest by testing its performance on various datasets.

I’m participating in SuperUROP because I haven’t had the opportunity to do in-depth research before. I wanted to pursue it seriously to try to get a sense of whether research is a path that I would like to continue on. I find my project exciting because I’m learning a lot not just about the research but about the field of data science as a whole which is really exciting.