mlsanityData to 14 Oct 2026

Evaluating Design Decisions for Dual Encoder-based Entity Disambiguation

arXiv 2505.11683 ↗First posted 16 May 2025cs.CLcited by 5

Not on the map. Why?

Abstract (arXiv)

Entity disambiguation (ED) is the task of linking mentions in text to corresponding entries in a knowledge base. Dual Encoders address this by embedding mentions and label candidates in a shared embedding space and applying a similarity metric to predict the correct label. In this work, we focus on evaluating key design decisions for Dual Encoder-based ED, such as its loss function, similarity metric, label verbalization format, and negative sampling strategy. We present the resulting model VerbalizED, a document-level Dual Encoder model that includes contextual label verbalizations and efficient hard negative sampling. Additionally, we explore an iterative prediction variant that aims to improve the disambiguation of challenging data points. Comprehensive experiments on AIDA-Yago validate the effectiveness of our approach, offering insights into impactful design choices that result in a new State-of-the-Art system on the ZELDA benchmark.

Key facts. Cited by 5 papers in the map (data through 14 Oct 2026). Builds on 22 of its 39 references that are in the map. First posted on arXiv 16 May 2025.

Builds on

22 of 39 references.

References · 22

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2020Autoregressive Entity Retrievalcited by 334
  3. 2019Learning Dense Representations for Entity Retrievalcited by 122
  4. 2017Deep Joint Entity Disambiguation with Local Neural Attentioncited by 79
  5. 2022ReFinED: An Efficient Zero-shot-capable Approach to End-to-End Entity Linkingcited by 74
  6. 2018Improving Entity Linking by Modeling Latent Relations between Mentionscited by 53
  7. 2020Investigating Entity Knowledge in BERT with Simple Neural End-To-End Entity Linkingcited by 49
  8. 2016Joint Learning of the Embedding of Words and Entities for Named Entity Disambiguationcited by 48
  9. 2021Evaluating Entity Disambiguation and the Role of Popularity in Retrieval-Based NLPcited by 31
  10. 2019Learning Dynamic Context Augmentation for Global Entity Linkingcited by 19
  11. 2020Empirical Evaluation of Pretraining Strategies for Supervised Entity Linkingcited by 17
  12. 2018Collective Entity Disambiguation with Structured Gradient Tree Boostingcited by 16
10 more
  1. 2024ReLiK: Retrieve and LinK, Fast and Accurate Entity Linking and Relation Extraction on an Academic Budgetcited by 16
  2. 2023SpEL: Structured Prediction for Entity Linkingcited by 11
  3. 2019Global Entity Disambiguation with BERTcited by 10
  4. 2024OneNet: A Fine-Tuning Free Framework for Few-Shot Entity Linking via Large Language Model Promptingcited by 9
  5. 2023A Read-and-Select Framework for Zero-shot Entity Linkingcited by 6
  6. 2024Entity Disambiguation via Fusion Entity Decodingcited by 4
  7. 2023Coherent Entity Disambiguation via Modeling Topic and Categorical Dependencycited by 3
  8. 2024Unified Examination of Entity Linking in Absence of Candidate Setscited by 3
  9. 2023Polar Ducks and Where to Find Them: Enhancing Entity Linking with Duck Typing and Polar Box Embeddingscited by 2
  10. 2024REXEL: An End-to-end Model for Document-Level Relation Extraction and Entity Linkingcited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 5.

5 papers.

  1. 2026LELA: an LLM-based Entity Linking Approach with Zero-Shot Domain Adaptationcited by 6
  2. 2026Semantic Reranking at Inference Time for Hard Examples in Rhetorical Role Labelingcited by 0
  3. 2026SPIEval: Evaluating Large Language Models as Mobile Assistants over Scattered Personal Informationcited by 0
  4. 2026Select, Don't Train: The Benefits of Modular Entity Disambiguation with LLM-Based Selectioncited by 0
  5. 2026From Location Phrases to Geographic Entities: Task-Adapted Retrieval for People Searchcited by 0