Taxonomy Completion via Triplet Matching Network (2021-01-06T00:00:00.000000Z)

TL;DR

Triplet Matching Network (TMN) is proposed, to find the appropriate pairs for a given query concept, and achieves the best performance on both taxonomy completion task and the previous taxonomy expansion task, outperforming existing methods.

Abstract

Automatically constructing taxonomy finds many applications in e-commerce and web search. One critical challenge is as data and business scope grow in real applications, new concepts are emerging and needed to be added to the existing taxonomy. Previous approaches focus on the taxonomy expansion, i.e. finding an appropriate hypernym concept from the taxonomy for a new query concept. In this paper, we formulate a new task, “taxonomy completion”, by discovering both the hypernym and hyponym concepts for a query. We propose Triplet Matching Network (TMN), to find the appropriate pairs for a given query concept. TMN consists of one primal scorer and multiple auxiliary scorers. These auxiliary scorers capture various fine-grained signals (e.g., query to hypernym or query to hyponym semantics), and the primal scorer makes a holistic prediction on triplet based on the internal feature representations of all auxiliary scorers. Also, an innovative channel-wise gating mechanism that retains task-specific information in concept representations is introduced to further boost model performance. Experiments on four real-world large-scale datasets show that TMN achieves the best performance on both taxonomy completion task and the previous taxonomy expansion task, outperforming existing methods.

Authors

Lei Li

15 papers

Yuning Mao

8 papers

Jieyu Zhang

3 papers

TL;DR

Abstract

Authors

References53 items

Co-Embedding Network Nodes and Hierarchical Labels with Taxonomy Based Generative Adversarial Networks

Self-supervised Auxiliary Learning with Meta-paths for Heterogeneous Graphs

Octet: Online Catalog Taxonomy Enrichment with Self-Supervision

STEAM: Self-Supervised Taxonomy Expansion with Mini-Paths

Expanding Taxonomies with Implicit Edge Semantics

TXtract: Taxonomy-Aware Knowledge Extraction for Thousands of Product Categories

TaxoExpan: Self-supervised Taxonomy Expansion with Position-Enhanced Graph Neural Network

Improving Semantic Analysis on Point Clouds via Auxiliary Supervision of Local Geometric Priors

Similarity Based Auxiliary Classifier for Named Entity Recognition

Hierarchical Text Classification with Reinforced Label Assignment

Every Child Should Have Parents: A Taxonomy Refinement Algorithm Based on Hyperbolic Term Embeddings

A User-Centered Concept Mining System for Query and Document Understanding at Tencent

Which Tasks Should Be Learned Together in Multi-task Learning?

Taxonomy-Aware Multi-Hop Reasoning Networks for Sequential Recommendation

Self-Supervised Generalisation with Meta Auxiliary Learning

Multi-Task Learning for Email Search Ranking with Auxiliary Query Clustering

TaxoGen: Constructing Topical Concept Taxonomy by Adaptive Term Embedding and Clustering

HiExpan: Task-Guided Taxonomy Construction by Hierarchical Tree Expansion

Representation Learning with Contrastive Predictive Coding

Enriching Taxonomies With Functional Domain Knowledge

Graph Convolutional Neural Networks for Web-Scale Recommender Systems

End-to-End Reinforcement Learning for Automatic Taxonomy Induction

Entity Set Search of Scientific Literature: An Unsupervised Ranking Approach

Junction Tree Variational Autoencoder for Molecular Graph Generation

Learning Longer-term Dependencies in RNNs with Auxiliary Losses

An Overview of Multi-Task Learning in Deep Neural Networks

Multitask Learning with Low-Level Auxiliary Tasks for Encoder-Decoder Based Speech Recognition

MetaPAD: Meta Pattern Discovery from Massive Text Corpora

Understand Short Texts by Harvesting and Analyzing Semantic Knowledge

Reinforcement Learning with Unsupervised Auxiliary Tasks

Learning Term Embeddings for Taxonomic Relation Identification Using Dynamic Weighting Neural Network

Conditional Image Synthesis with Auxiliary Classifier GANs

SemEval-2016 Task 14: Semantic Taxonomy Enrichment

An Overview of Microsoft Academic Service (MAS) and Applications

Inclusive yet Selective: Supervised Distributional Hypernymy Detection

Taxonomy discovery for personalized recommendation

Reasoning With Neural Tensor Networks for Knowledge Base Completion

A latent factor model for highly multi-relational data

PATTY: A Taxonomy of Relational Patterns with Semantic Types

Constructing Task-Specific Taxonomies for Document Collection Browsing

Probase: a probabilistic taxonomy for text understanding

Wikidata: a new platform for collaborative data collection

Modelling Relational Data using Bayesian Clustered Tensor Factorization

Characterising Measures of Lexical Distributional Similarity

Medical Subject Headings (MeSH).

Snowball: extracting relations from large plain-text collections

Automatic Acquisition of Hyponyms from Large Text Corpora

One state-of-the-art taxonomy expansion framework which leverages position-enhanced graph neural network to capture local information and InfoNCE loss

Adaptive Auxiliary Task Weighting for Reinforcement Learning

It incorporates Single Layer Model with a bilinear tensor layer that directly relates the two concept embeddings across multiple dimensions and a bias vector

Bilinear Model (Sutskever, Salakhutdinov, and Tenen-baum

Single Layer Model : A model that scores tuple by a standard single layer neural network which inputs the concatenation of the concept embeddings

Multiple Layer Model : An extension of Single Layer Model that replaces the single layer neural network with multiple layer neural network

Field of Study

Venue Information

Name

Type

URL

Alternate Names