A Survey of Event Causality Identification: Principles, Taxonomy, Challenges, and Assessment

2024-11-15

arXiv.org

AI Summary

Event Causal Identification (ECI) suffers from fragmented methodologies, inconsistent evaluation protocols, and poor cross-lingual and zero-shot generalization. To address these challenges, this work introduces the first unified dichotomous classification framework—Sentence-level ECI (SECI) and Document-level ECI (DECI)—establishing a coherent conceptual taxonomy and standardized benchmark. We conduct a systematic, multi-paradigm quantitative evaluation across six mainstream approaches: feature matching, semantic encoding, causal pretraining, prompt-based fine-tuning, knowledge integration, and event graph neural networks. Our analysis characterizes the interpretability, long-range dependency modeling capacity, and low-resource robustness of each paradigm. Experiments span two major benchmarks, revealing performance ceilings and prevalent error patterns. We identify three critical frontiers for future advancement: interpretable causal reasoning, cross-sentence long-range dependency modeling, and low-resource generalization.

Technology Category

• Reasoning under Uncertainty: Causality
• Machine Learning: Causal Learning
• Knowledge Representation and Reasoning: Action, Change, and Causality

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Event Causality Identification (ECI) has become a crucial task in Natural Language Processing (NLP), aimed at automatically extracting causalities from textual data. In this survey, we systematically address the foundational principles, technical frameworks, and challenges of ECI, offering a comprehensive taxonomy to categorize and clarify current research methodologies, as well as a quantitative assessment of existing models. We first establish a conceptual framework for ECI, outlining key definitions, problem formulations, and evaluation standards. Our taxonomy classifies ECI methods according to the two primary tasks of sentence-level (SECI) and document-level (DECI) event causality identification. For SECI, we examine feature pattern-based matching, deep semantic encoding, causal knowledge pre-training and prompt-based fine-tuning, and external knowledge enhancement methods. For DECI, we highlight approaches focused on event graph reasoning and prompt-based techniques to address the complexity of cross-sentence causal inference. Additionally, we analyze the strengths, limitations, and open challenges of each approach. We further conduct an extensive quantitative evaluation of various ECI methods on two benchmark datasets. Finally, we explore future research directions, highlighting promising pathways to overcome current limitations and broaden ECI applications.

Problem

• Surveying Event Causality Identification (ECI) methods in NLP
• Classifying ECI techniques into SECI and DECI frameworks
• Evaluating challenges and future directions in ECI research

Innovation

• Classification framework for ECI methods
• Deep semantic encoding techniques
• Prompt-based fine-tuning with LLMs