Semi-supervised Multi-sensor Classification via Consensus-based Multi-View Maximum Entropy Discrimination
Xie, Nasrabadi, Hero · cs.IT,cs.AI,cs.LG,math.IT · 2015-07-05 · 原文
In this paper, we consider multi-sensor classification when there is a large number of unlabeled samples. The problem is formulated under the multi-view learning framework and a Consensus-based Multi-View Maximum Entropy Discrimination (CMV-MED) algorithm is proposed. By iteratively maximizing the stochastic agreement between multiple classifiers on the unlabeled dataset, the algorithm simultaneously learns multiple high accuracy classifiers. We demonstrate that our proposed method can yield improved performance over previous multi-view learning approaches by comparing performance on three real multi-sensor data sets.
讲义
讲义·推断 依据「原文」自动生成的结构化摘要（推断），非原文表述；以原文为准。
1. 人话版
In this paper, we consider multi-sensor classification when there is a large number of unlabeled samples.
The problem is formulated under the multi-view learning framework and a Consensus-based Multi-View Maximum Entropy Discrimination (CMV-MED) algorithm is proposed.
2. 领域脉络
本文类目：cs.IT、cs.AI、cs.LG、math.IT，属于其所在研究脉络的最新进展。
3. 机制拆解
By iteratively maximizing the stochastic agreement between multiple classifiers on the unlabeled dataset, the algorithm simultaneously learns multiple high accuracy classifiers.
We demonstrate that our proposed method can yield improved performance over previous multi-view learning approaches by comparing performance on three real multi-sensor data sets.
4. 证据与数字
摘要未给出量化结果——留意原文的实验与数据。
5. 反例与边界
摘要未声明局限与反例——这是需要警惕的信号，精读时先问边界。
6. 跨领域连接与意外收获
横跨 4 个类目（cs.IT、cs.AI、cs.LG、math.IT），关注其在你兴趣板块间的迁移面。
7. 可复用方法
把本文机制与你手头项目对照，找一个两周内能验证的最小实验。
8. 术语表
精读时把不熟的术语记入此处，作为下次回忆的锚点。