概率条件随机场的损失敏感训练

Jul, 2011

Loss-sensitive Training of Probabilistic Conditional Random Fields

Maksims N. Volkovs, Hugo Larochelle, Richard S. Zemel

TL;DR讨论了训练概率条件随机场中最大似然以外的代价函数，该方法以对比排序为基准的实验验证了将代价函数信息纳入到概率训练中的重要性，其中代价启发KL目标函数的表现最优。

Abstract

We consider the problem of training probabilistic conditional random fields (CRFs) in the context of a task where performance is measured using a specific loss function. While →