全部

Probabilistic Dimensionality Reduction via Structure Learning

  • 演讲者:Wang Li(德克萨斯大学阿灵顿分校)

  • 时间:2019-06-26 17:00-18:00

  • 地点:慧园3栋 415报告厅

Abstract: We propose an alternative probabilistic dimensionality reduction framework that can naturally integrate the generative model and the locality information of data. Based on this framework, we present a new model, which is able to learn a set of embedding points in a low-dimensional space by retaining the inherent structure from high-dimensional data. The objective function of this new model can be equivalently interpreted as two coupled learning problems, i.e., structure learning and the learning of projection matrix. Inspired by this interesting interpretation, we propose another model, which finds a set of embedding points that can directly form an explicit graph structure. We proved that the model by learning explicit graphs generalizes the reversed graph embedding method, but leads to a natural interpretation from Bayesian perspective. This can greatly facilitate data visualization and scientific discovery in downstream analysis.