Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Retrieval-Enhanced Named Entity Recognition

Enzo Shiraishi, Raphael Y. de Camargo, Henrique L. P. Silva, Ronaldo C. Prati

TL;DR

RENER (Retrieval-Enhanced Named Entity Recognition), a technique for named entity recognition using autoregressive language models based on In-Context Learning and information retrieval techniques, which is modular and independent of the underlying language model and information retrieval algorithms.

Abstract

When combined with In-Context Learning, a technique that enables models to adapt to new tasks by incorporating task-specific examples or demonstrations directly within the input prompt, autoregressive language models have achieved good performance in a wide range of tasks and applications. However, this combination has not been properly explored in the context of named entity recognition, where the structure of this task poses unique challenges. We propose RENER (Retrieval-Enhanced Named Entity Recognition), a technique for named entity recognition using autoregressive language models based on In-Context Learning and information retrieval techniques. When presented with an input text, RENER fetches similar examples from a dataset of training examples that are used to enhance a language model to recognize named entities from this input text. RENER is modular and independent of the underlying language model and information retrieval algorithms. Experimental results show that in the CrossNER collection we achieve state-of-the-art performance with the proposed technique and that information retrieval can increase the F-score by up to 11 percentage points.

Retrieval-Enhanced Named Entity Recognition

TL;DR

RENER (Retrieval-Enhanced Named Entity Recognition), a technique for named entity recognition using autoregressive language models based on In-Context Learning and information retrieval techniques, which is modular and independent of the underlying language model and information retrieval algorithms.

Abstract

When combined with In-Context Learning, a technique that enables models to adapt to new tasks by incorporating task-specific examples or demonstrations directly within the input prompt, autoregressive language models have achieved good performance in a wide range of tasks and applications. However, this combination has not been properly explored in the context of named entity recognition, where the structure of this task poses unique challenges. We propose RENER (Retrieval-Enhanced Named Entity Recognition), a technique for named entity recognition using autoregressive language models based on In-Context Learning and information retrieval techniques. When presented with an input text, RENER fetches similar examples from a dataset of training examples that are used to enhance a language model to recognize named entities from this input text. RENER is modular and independent of the underlying language model and information retrieval algorithms. Experimental results show that in the CrossNER collection we achieve state-of-the-art performance with the proposed technique and that information retrieval can increase the F-score by up to 11 percentage points.

Paper Structure

This paper contains 20 sections, 6 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. In-Context Learning
  4. Information Retrieval
  5. Named Entity Recognition
  6. RENER
  7. Retrieval module
  8. BM25
  9. Semantic search
  10. Maximal Marginal Relevance
  11. Reciprocal Rank Fusion
  12. Recognition module
  13. Evaluation
  14. Experiments
  15. Data
  16. ...and 5 more sections

Figures (6)

  • Figure 1: Flowchart representing the retrieval and recognition modules in the inference process.
  • Figure 2: Example illustrations and their equivalent representations in YAML, based on samples from the dataset of the CoNLL 2003 shared task.
  • Figure 3: Flowchart representing the inference steps on the retrieval module.
  • Figure 4: Flowchart representing the inference steps on the recognition module.
  • Figure 5: Chart relating the F-score and $k$ on the validation datasets of CrossNER for all retrieval mechanisms using Gemini.
  • ...and 1 more figures