STEAM: Self-Supervised Taxonomy Expansion with Mini-Paths

  title={STEAM: Self-Supervised Taxonomy Expansion with Mini-Paths},
  author={Yue Yu and Yinghao Li and Jiaming Shen and Haoyang Feng and Jimeng Sun and Chao Zhang},
  journal={Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery \& Data Mining},
  • Yue YuYinghao Li Chao Zhang
  • Published 18 June 2020
  • Computer Science
  • Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining
Taxonomies are important knowledge ontologies that underpin numerous applications on a daily basis, but many taxonomies used in practice suffer from the low coverage issue. We study the taxonomy expansion problem, which aims to expand existing taxonomies with new concept terms. We propose a self-supervised taxonomy expansion model named STEAM, which leverages natural supervision in the existing taxonomy for expansion. To generate natural self-supervision signals, STEAM samples mini-paths from… 

Figures and Tables from this paper

TEMP: Taxonomy Expansion with Dynamic Margin Loss through Taxonomy-Paths

This paper proposes TEMP, a self-supervised taxonomy expansion method, which predicts the position of new concepts by ranking the generated taxonomy-paths, and employs pre-trained contextual encoders in taxonomy construction and hypernym detection problems.

TaxoEnrich: Self-Supervised Taxonomy Completion via Structure-Semantic Representations

Extensive experiments on four large real-world datasets from different domains show that TaxoEnrich achieves the best performance among all evaluation metrics and outperforms previous state-of-the-art methods by a large margin.

Enquire One’s Parent and Child Before Decision: Fully Exploit Hierarchical Structure for Self-Supervised Taxonomy Expansion

By better exploiting the hierarchical structure and optimizing taxonomy’s coherence, HEF vastly surpasses the prior state-of-the-art on three benchmark datasets by an average improvement of 46.7% in accuracy and 32.3% in mean reciprocal rank.

QEN: Applicable Taxonomy Completion via Evaluating Full Taxonomic Relations

The Quadruple Evaluation Network (QEN) is proposed, a novel taxonomy completion framework that utilizes easily accessible term descriptions as input, and applies pretrained language model and code attention for accurate inference while reducing online computation.

TaxoPrompt: A Prompt-based Generation Method with Taxonomic Context for Self-Supervised Taxonomy Expansion

This paper proposes TaxoPrompt, a framework that learns the global structure by prompt tuning by exploiting the taxonomic context constructed by a variant of the random walk algorithm.

Learning What You Need from What You Did: Product Taxonomy Expansion with User Behaviors Supervision

This paper presents a self-supervised and user behavior-oriented product taxonomy expansion framework to append new concepts into existing taxonomies automatically and constructs a high-quality and balanced training dataset from existing taxonomy with no supervision.

Enhancing Taxonomy Completion with Concept Generation via Fusing Relational Representations

GenTaxo is proposed to enhance taxonomy completion by identifying positions in existing taxonomyies that need new concepts and then generating appropriate concept names, and improves the completeness of taxonomies over existing methods.

TEAM: A multitask learning based Taxonomy Expansion approach for Attach and Merge

This paper proposes a novel multi-task learning-based deep learning method known as Taxonomy Expansion with Attach and Merge (TEAM) that performs both the merge and attach operations.

Who Should Go First? A Self-Supervised Concept Sorting Model for Improving Taxonomy Expansion

TaxoOrder is a novel self-supervised framework that simultaneously discovers hypernym-hyponym relations among new concepts and decides their insertion order, leading to better-resulting taxonomies with comparison to baselines under various evaluation metrics.

HyperExpan: Taxonomy Expansion with Hyperbolic Representation Learning

A taxonomy expansion algorithm that seeks to preserve the structure of a taxonomy in a more expressive hyperbolic embedding space and learn to represent concepts and their relations with a Hyperbolic Graph Neural Network (HGNN).



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

A novel self-supervised framework, named TaxoExpan, which automatically generates a set of ⟨query concept, anchor concept⟩ pairs from the existing taxonomy as training data, and develops two innovative techniques, including a position-enhanced graph neural network that encodes the local structure of an anchor concept in theexisting taxonomy and a noise-robust training objective that enables the learned model to be insensitive to the label noise in the self- supervision data.

Expanding Taxonomies with Implicit Edge Semantics

This work proposes Arborist, an approach to automatically expand textual taxonomies by predicting the parents of new taxonomy nodes, and proposes a principled formulation of the margin function, which theoretically guarantees that Arborist minimizes an upper-bound on the shortest-path distance between the predicted parents and actual parents in the taxonomy.

End-to-End Reinforcement Learning for Automatic Taxonomy Induction

This work presents a novel end-to-end reinforcement learning approach to automatic taxonomy induction from a set of terms that outperforms prior state-of-the-artTaxonomy induction methods up to 19.6% on ancestor F1.

Taxonomy Construction of Unseen Domains via Graph-based Cross-Domain Knowledge Transfer

This paper proposes Graph2Taxo, a GNN-based cross-domain transfer framework for the taxonomy construction task, which uses a noisy graph constructed from automatically extracted noisy hyponym hypernym candidate pairs, and a novel method of directed acyclic graph (DAG) generation for taxonomyConstruction.

Structured Learning for Taxonomy Induction with Belief Propagation

This model incorporates heterogeneous relational evidence about both hypernymy and siblinghood, captured by semantic features based on patterns and statistics from Web n-grams and Wikipedia abstracts, and uses loopy belief propagation along with a directed spanning tree algorithm for the core hypernym factor.

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

This paper proposes TXtract, a taxonomy-aware knowledge extraction model that applies to thousands of product categories organized in a hierarchical taxonomy that is both scalable, as it trains a single model for thousands of categories, and effective, asIt extracts category-specific attribute values.

Distributional Inclusion Vector Embedding for Unsupervised Hypernymy Detection

Distributional inclusion vector embedding (DIVE) is introduced, a simple-to-implement unsupervised method of hypernym discovery via per-word non-negative vector embeddings which preserve the inclusion property of word contexts.

Learning Concept Taxonomies from Multi-modal Data

This work proposes a probabilistic model for taxonomy induction by jointly leveraging text and images and designs end-to-end features based on distributed representations of images and words to avoid hand-crafted feature engineering.

Taxonomy Induction Using Hypernym Subsequences

This work proposes a novel, semi-supervised approach towards domain taxonomy induction from an input vocabulary of seed terms that outperforms state-of-the-artTaxonomy induction approaches across four languages and is robust to the presence of noise in the input vocabulary.

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

We introduce the use of Poincaré embeddings to improve existing state-of-the-art approaches to domain-specific taxonomy induction from text as a signal for both relocating wrong hyponym terms within