Who Is the Story About? Protagonist Entity Recognition in News

2025-11-10

AI Summary

Traditional named entity recognition (NER) treats all entities in news texts uniformly, failing to identify core organizational entities that drive narrative progression—termed “protagonist entities”—thereby limiting modeling of event salience and narrative focus. Method: We formally define, manually annotate, and validate the novel Protagonist Entity Recognition (PER) task. To overcome data scarcity, we propose an NER-guided large language model (LLM) prompting framework that efficiently generates high-quality supervision signals. Model evaluation is conducted on an expert-constructed gold-standard corpus under constrained-context settings. Contribution/Results: PER is demonstrated to be a feasible, semantically well-defined information extraction direction. LLMs fine-tuned via our guidance achieve high consistency with human judgments of narrative importance, significantly outperforming conventional NER baselines. This work establishes a new paradigm for narrative understanding and event-centric analysis in computational journalism and social science applications.

Technology Category

• Natural Language Processing: Information Extraction
• Machine Learning: Large Multimodal Models (LMMs)
• Humans and AI: Game Design — Procedural Content Generation & Storytelling

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Large language models for search
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

News articles often reference numerous organizations, but traditional Named Entity Recognition (NER) treats all mentions equally, obscuring which entities genuinely drive the narrative. This limits downstream tasks that rely on understanding event salience, influence, or narrative focus. We introduce Protagonist Entity Recognition (PER), a task that identifies the organizations that anchor a news story and shape its main developments. To validate PER, we compare he predictions of Large Language Models (LLMs) against annotations from four expert annotators over a gold corpus, establishing both inter-annotator consistency and human-LLM agreement. Leveraging these findings, we use state-of-the-art LLMs to automatically label large-scale news collections through NER-guided prompting, generating scalable, high-quality supervision. We then evaluate whether other LLMs, given reduced context and without explicit candidate guidance, can still infer the correct protagonists. Our results demonstrate that PER is a feasible and meaningful extension to narrative-centered information extraction, and that guided LLMs can approximate human judgments of narrative importance at scale.

Problem

• Identifying key organizations driving news narratives
• Extending NER to capture narrative centrality in stories
• Evaluating LLMs' ability to recognize protagonist entities

Innovation

• Introducing Protagonist Entity Recognition for news
• Using LLMs with NER-guided prompting for labeling
• Evaluating LLMs' ability to infer protagonists automatically