论文阅读 [TPAMI

论文阅读 [TPAMI

论文阅读 [TPAMI-2022] Unsupervised Heterogeneous Coupling Learning for Categorical Representation

论文搜索(studyai)

搜索论文: Unsupervised Heterogeneous Coupling Learning for Categorical Representation

搜索论文: /?q=Unsupervised+Heterogeneous+Coupling+Learning+for+Categorical+Representation

关键字(Keywords)

Couplings; Kernel; Frequency measurement; Complexity theory; Task analysis; Shape; Image color analysis; Coupling learning; heterogeneity learning; Non-IID learning; representation learning; similarity learning; categorical data; categorical data representation; unsupe

机器学习; 自然语言处理

无监督学习; 语言表示学习

摘要(Abstract)

Complex categorical data is often hierarchically coupled with heterogeneous relationships between attributes and attribute values and the couplings between objects.

复杂的分类数据通常在层次上与属性和属性值之间的异构关系以及对象之间的耦合相耦合。.

Such value-to-object couplings are heterogeneous with complementary and inconsistent interactions and distributions.

这种价值到对象的耦合是异质的，具有互补和不一致的交互和分布。.

Limited research exists on unlabeled categorical data representations, ignores the heterogeneous and hierarchical couplings, underestimates data characteristics and complexities, and overuses redundant information, etc.

对未标记的分类数据表示的研究有限，忽略了异构和分层耦合，低估了数据的特征和复杂性，过度使用了冗余信息等。.

The deep representation learning of unlabeled categorical data is challenging, overseeing such value-to-object couplings, complementarity and inconsistency, and requiring large data, disentanglement, and high computational power.

未标记分类数据的深度表示学习具有挑战性，需要监督此类值到对象的耦合、互补性和不一致性，并且需要大数据、解纠缠和高计算能力。.

This work introduces a shallow but powerful UNsupervised heTerogeneous couplIng lEarning (UNTIE) approach for representing coupled categorical data by untying the interactions between couplings and revealing heterogeneous distributions embedded in each type of couplings.

这项工作引入了一种肤浅但功能强大的无监督异构耦合学习（UNTE）方法，通过解开耦合之间的交互并揭示嵌入每种类型耦合中的异构分布来表示耦合的分类数据。.

UNTIE is efficiently

UNTE是高效优化的w.r.t。.

a kernel k-means objective function for unsupervised representation learning of heterogeneous and hierarchical value-to-object couplings.

一个核k-均值目标函数，用于异构和分层值到对象耦合的无监督表示学习。.

Theoretical analysis shows that UNTIE can represent categorical data with maximal separability while effectively represent heterogeneous couplings and disclose their roles in categorical data.

理论分析表明，UNTE能够以最大可分性表示分类数据，同时有效地表示异构耦合，揭示它们在分类数据中的作用。.

The UNTIE-learned representations make significant performance improvement against the state-of-the-art categorical representations and deep representation models on 25 categorical data sets with diversified characteristics…

在25个具有多种特征的分类数据集上，与最先进的分类表示和深度表示模型相比，联合学习表示法的性能显著提高。。.

作者(Authors)

[‘Chengzhang Zhu’, ‘Longbing Cao’, ‘Jianping Yin’]