学术科研
research
we are in the exciting era where ai and database technologies are colliding and merging. in this talk, i will introduce the motivations and several opportunities to integrate machine learning/deep learning and database research, in both directions. as a case study, i will focus on a new deep learning-based model for range selectivity estimation for high-dimensional data. the model is guaranteed to be consistent, i.e., monotonic in the threshold parameter, and has demonstrated superior performance against state-of-the-art methods in all error measures across several datasets and distance/similarity functions.