TL;DR本文提出一种新的基于低秩线性回归和张量幂法的计算高效且具有保证一致性的鉴别性潜变量模型估计方法,该方法在混合线性回归模型中被验证其收敛性能优于局部优化 EM 算法。
Abstract
discriminative latent-variable models are typically learned using EM or
gradient-based optimization, which suffer from local optima. In this paper, we
develop a new computationally efficient and provably consiste