Reveal the Unknown: Out-of-Knowledge-Base Mention Discovery with Entity Linking (2023-02-14T00:00:00.000000Z)

TL;DR

BLINKout is proposed, a new BERT-based Entity Linking (EL) method which can identify mentions that do not have corresponding KB entities by matching them to a special NIL entity, with synonym enhancement.

Abstract

Discovering entity mentions that are out of a Knowledge Base (KB) from texts plays a critical role in KB maintenance, but has not yet been fully explored. The current methods are mostly limited to the simple threshold-based approach and feature-based classification, and the datasets for evaluation are relatively rare. We propose BLINKout, a new BERT-based Entity Linking (EL) method which can identify mentions that do not have corresponding KB entities by matching them to a special NIL entity. To better utilize BERT, we propose new techniques including NIL entity representation and classification, with synonym enhancement. We also apply KB Pruning and Versioning strategies to automatically construct out-of-KB datasets from common in-KB EL datasets. Results on five datasets of clinical notes, biomedical publications, and Wikipedia articles in various domains show the advantages of BLINKout over existing methods to identify out-of-KB mentions for the medical ontologies, UMLS, SNOMED CT, and the general KB, WikiData.

Authors

Yuan He

5 papers

Jiaoyan Chen

17 papers

Hang Dong

4 papers

TL;DR

Abstract

Authors

References64 items

DeepOnto: A Python Package for Ontology Engineering with Deep Learning

NASTyLinker: NIL-Aware Scalable Transformer-based Entity Linker

NILK: Entity Linking Dataset Targeting NIL-linking Cases

ReFinED: An Efficient Zero-shot-capable Approach to End-to-End Entity Linking

NILINKER: Attention-based approach to NIL Entity Linking

EDIN: An End-to-end Benchmark and Pipeline for Unknown Entity Discovery and Indexing

Machine Learning-Friendly Biomedical Datasets for Equivalence and Subsumption Ontology Matching

A simple neural vector space model for medical concept normalization using concept embeddings

KG-ZESHEL: Knowledge Graph-Enhanced Zero-Shot Entity Linking

Entity Linking Meets Deep Learning: Techniques and Solutions

SimCSE: Simple Contrastive Learning of Sentence Embeddings

Fast and Effective Biomedical Entity Linking Using a Dual Encoder

Read, Retrospect, Select: An MRC Framework to Short Text Entity Linking

A Lightweight Neural Model for Biomedical Entity Linking

Self-Alignment Pretraining for Biomedical Entity Representations

Extended Overview of CLEF HIPE 2020: Named Entity Processing on Historical Newspapers

COMETA: A Corpus for Medical Entity Linking in the Social Media

PyTorch Metric Learning

Domain-Specific Language Model Pretraining for Biomedical Natural Language Processing

Neural Entity Linking: A Survey of Models based on Deep Learning

Novel Entity Discovery from Web Tables

Scalable Zero-shot Entity Linking with Dense Entity Retrieval

Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks

BERT-based Ranking for Biomedical Entity Normalization

Joint Learning of Named Entity Recognition and Entity Linking

Transfer Learning in Biomedical Natural Language Processing: An Evaluation of BERT and ELMo on Ten Benchmarking Datasets

MedMentions: A Large Biomedical Corpus Annotated with UMLS Concepts

BioBERT: a pre-trained biomedical language representation model for biomedical text mining

Advances in Pre-Training Distributed Word Representations

Decoupled Weight Decay Regularization

Gastrointestinal disorders in Curry–Jones syndrome: Clinical and molecular insights from an affected newborn

Lessons learnt from the Named Entity rEcognition and Linking (NEEL) challenge series

One for All: Towards Language Independent Named Entity Linking

A Recurrent Mosaic Mutation in SMO, Encoding the Hedgehog Signal Transducer Smoothened, Is the Major Cause of Curry-Jones Syndrome.

The Knowledge Awakens: Keeping Knowledge Bases Fresh with Emerging Entities

Exploring Multiple Feature Spaces for Novel Entity Discovery

Sieve-Based Entity Linking for the Biomedical Domain

Entity Linking with a Knowledge Base: Issues, Techniques, and Solutions

Wikidata

Overview of the ShARe/CLEF eHealth Evaluation Lab 2014

Discovering emerging entities with ambiguous names

Entity Disambiguation for Knowledge Base Population

HLTCOE Approaches to Knowledge Base Population at TAC 2009

Using Encyclopedic Knowledge for Named entity Disambiguation

Toward principles for the design of ontologies used for knowledge sharing?

Knowledge Graph Population with Out-of-KG Entities

Research and Applications The 2019 n2c2/UMass Lowell shared task on clinical concept normalization

Association for Computational Linguistics, Online and Punta Cana

Entity Linking and Discovery via Arborescence-based Supervised Clustering

2020) to use a maximum of 32 tokens

BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

WWW '16 Companion). International World Wide Web Conferences Steering Committee

Entity Linking: Finding Extracted Entities in a Knowledge Base

ShARe- CLEF eHealth 2013: Natural Language Processing and Information Retrieval for Clinical Care 1.0. https://physionet.org/content/shareclefehealth2013/1.0/ Task1ShAReGuidelines2013.pdf

Overview of the TAC 2010 Knowledge Base Population Track

domain-specific in

SNOMED-CT: The advanced terminology and coding system for eHealth.

The Unified Medical Language System (UMLS): integrating biomedical terminology

Chap-man

Conference acronym ’XX, June 03–05, 2018

World Health Organization. 2022. Tracking SARS-CoV-2 variants

2022. Tracking

Multimodality, and Visualization - 4th International Conference of the CLEF Initiative, CLEF

Reveal the Unknown: Out-of-Knowledge-Base Mention

Field of Study

Journal Information

Name

Venue Information

Name

Type

URL

Alternate Names