# Bootleg: Chasing the Tail with Self-Supervised Named Entity Disambiguation

arXiv: https://arxiv.org/abs/2010.10363 · HTML: https://mlsanity.com/paper/2010.10363/

## Key facts

- Cited by 24 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 20 Oct 2020.

## Built on by (most cited first, 24 of 24)

- [Domino: Discovering Systematic Errors with Cross-Modal Embeddings](https://mlsanity.com/paper/2203.14960/) (2022)
- [Robustness Gym: Unifying the NLP Evaluation Landscape](https://mlsanity.com/paper/2101.04840/) (2021)
- [RnG-KBQA: Generation Augmented Iterative Ranking for Knowledge Base Question Answering](https://mlsanity.com/paper/2109.08678/) (2021)
- [ReFinED: An Efficient Zero-shot-capable Approach to End-to-End Entity Linking](https://mlsanity.com/paper/2207.04108/) (2022)
- [Evaluating Entity Disambiguation and the Role of Popularity in Retrieval-Based NLP](https://mlsanity.com/paper/2106.06830/) (2021)
- [Saga: A Platform for Continuous Construction and Serving of Knowledge At Scale](https://mlsanity.com/paper/2204.07309/) (2022)
- [Cross-Domain Data Integration for Named Entity Disambiguation in Biomedical Text](https://mlsanity.com/paper/2110.08228/) (2021)
- [TempEL: Linking Dynamically Evolving and Newly Emerging Entities](https://mlsanity.com/paper/2302.02500/) (2023)
- [On the Scaling Laws of Geographical Representation in Language Models](https://mlsanity.com/paper/2402.19406/) (2024)
- [Benchmarking Long-tail Generalization with Likelihood Splits](https://mlsanity.com/paper/2210.06799/) (2022)
- [TABi: Type-Aware Bi-Encoders for Open-Domain Entity Retrieval](https://mlsanity.com/paper/2204.08173/) (2022)
- [Multi-group Learning for Hierarchical Groups](https://mlsanity.com/paper/2402.00258/) (2024)
- [Ember: No-Code Context Enrichment via Similarity-Based Keyless Joins](https://mlsanity.com/paper/2106.01501/) (2021)
- [Improving Entity Disambiguation by Reasoning over a Knowledge Base](https://mlsanity.com/paper/2207.04106/) (2022)
- [Do I have the Knowledge to Answer? Investigating Answerability of Knowledge Base Questions](https://mlsanity.com/paper/2212.10189/) (2022)
- [Managing ML Pipelines: Feature Stores and the Coming Wave of Embedding Ecosystems](https://mlsanity.com/paper/2108.05053/) (2021)
- [Increasing Coverage and Precision of Textual Information in Multilingual Knowledge Graphs](https://mlsanity.com/paper/2311.15781/) (2023)
- [Entity Disambiguation via Fusion Entity Decoding](https://mlsanity.com/paper/2404.01626/) (2024)
- [Coherent Entity Disambiguation via Modeling Topic and Categorical Dependency](https://mlsanity.com/paper/2311.03253/) (2023)
- [Polar Ducks and Where to Find Them: Enhancing Entity Linking with Duck Typing and Polar Box Embeddings](https://mlsanity.com/paper/2305.12027/) (2023)
- [Ered: Enhanced Text Representations with Entities and Descriptions](https://mlsanity.com/paper/2208.08954/) (2022)
- [A Unified Knowledge Graph Augmentation Service for Boosting Domain-specific NLP Tasks](https://mlsanity.com/paper/2212.05251/) (2022)
- [Metadata Shaping: Natural Language Annotations for the Tail](https://mlsanity.com/paper/2110.08430/) (2021)
- [Iteratively Improving Biomedical Entity Linking and Event Extraction via Hard Expectation-Maximization](https://mlsanity.com/paper/2305.14645/) (2023)

## Abstract (arXiv)

A challenge for named entity disambiguation (NED), the task of mapping textual mentions to entities in a knowledge base, is how to disambiguate entities that appear rarely in the training data, termed tail entities. Humans use subtle reasoning patterns based on knowledge of entity facts, relations, and types to disambiguate unfamiliar entities. Inspired by these patterns, we introduce Bootleg, a self-supervised NED system that is explicitly grounded in reasoning patterns for disambiguation. We define core reasoning patterns for disambiguation, create a learning procedure to encourage the self-supervised model to learn the patterns, and show how to use weak supervision to enhance the signals in the training data. Encoding the reasoning patterns in a simple Transformer architecture, Bootleg meets or exceeds state-of-the-art on three NED benchmarks. We further show that the learned representations from Bootleg successfully transfer to other non-disambiguation tasks that require entity-based knowledge: we set a new state-of-the-art in the popular TACRED relation extraction task by 1.0 F1 points and demonstrate up to 8% performance lift in highly optimized production search and assistant tasks at a major technology company