公开(公告)号：US08977579B2

公开(公告)日：2015-03-10

申请号：US13649823

申请日：2012-10-11

Applicant: NEC Laboratories America, Inc.

Inventor： Yunlong He ,  Yanjun Qi ,  Koray Kavukcuoglu

IPC: G06F15/18 ,  G06N99/00 ,  A61M5/00 ,  A61M25/10 ,  A61M25/00

CPC classification number: G06N99/005 ,  A61M5/007 ,  A61M25/0026 ,  A61M25/10 ,  A61M25/1011 ,  A61M2025/105

Abstract: Disclosed is a general learning framework for computer implementation that induces sparsity on the undirected graphical model imposed on the vector of latent factors. A latent factor model SLFA is disclosed as a matrix factorization problem with a special regularization term that encourages collaborative reconstruction. Advantageously, the model may simultaneously learn the lower-dimensional representation for data and model the pairwise relationships between latent factors explicitly. An on-line learning algorithm is disclosed to make the model amenable to large-scale learning problems. Experimental results on two synthetic data and two real-world data sets demonstrate that pairwise relationships and latent factors learned by the model provide a more structured way of exploring high-dimensional data, and the learned representations achieve the state-of-the-art classification performance.

Abstract translation: 公开了一种用于计算机实现的通用学习框架，其在对潜在因素的向量施加的无向图形模型上引起稀疏性。 潜在因素模型SLFA被公开为具有鼓励协作重建的特殊正则化术语的矩阵分解问题。 有利地，该模型可以同时学习数据的低维表示，并明确地模拟潜在因素之间的成对关系。 公开了一种在线学习算法，使得该模型适合于大规模学习问题。 两个合成数据和两个现实世界数据集的实验结果表明，该模型获得的成对关系和潜在因素提供了一种更结构化的探索高维数据的方法，并且学习的表示实现了最先进的分类 性能。