In-context Contrastive Learning for Event Causality Identification

TL;DR

An ICCL model is proposed that utilizes contrastive learning to enhance the effectiveness of both positive and negative demonstrations and applies contrastive learning to event pairs to better facilitate event causality identification.

Abstract

Event Causality Identification (ECI) aims at determining the existence of a causal relation between two events. Although recent prompt learning-based approaches have shown promising improvements on the ECI task, their performance are often subject to the delicate design of multiple prompts and the positive correlations between the main task and derivate tasks. The in-context learning paradigm provides explicit guidance for label prediction in the prompt learning paradigm, alleviating its reliance on complex prompts and derivative tasks. However, it does not distinguish between positive and negative demonstrations for analogy learning. Motivated from such considerations, this paper proposes an In-Context Contrastive Learning (ICCL) model that utilizes contrastive learning to enhance the effectiveness of both positive and negative demonstrations. Additionally, we apply contrastive learning to event pairs to better facilitate event causality identification. Our ICCL is evaluated on the widely used corpora, including the EventStoryLine and Causal-TimeBank, and results show significant performance improvements over the state-of-the-art algorithms.

Figures (7)

• Figure 1: Illustration of our motivation. The event pairs are highlighted in different colors.
• Figure 2: Illustration of our ICCL framwork.
• Figure 3: Comparision of ICCL and In-context model when using differenr numbers of causal and non-causal demonstrations on ESC corpus.
• Figure 4: Results of few shot on ESC corpus. We replicated ERGO and get its few-shot results in the figure.
• Figure 5: Visualization of the event pairs' embedding encoded by different models on ESC corpus