Ya Su, Hu Zhang, Dan Qiao, YuJie Wang, Yunxiao Zhao, Yue Fan, Shike Li, Ru Li, Hongye Tan. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Event Causality Identification (ECI) aims to identify fine-grained causal relationships between events in an unstructured text. Existing ECI methods primarily rely on knowledge enhanced and graph-based reasoning approaches, but they often overlook the dependencies between similar events. Additionally, the connection between unstructured text and structured knowledge is relatively weak. Therefore, this paper proposes an ECI method enhanced by LLM Knowledge and Concept-Level Event Relations (LKCER). Specifically, LKCER constructs a conceptual-level heterogeneous event graph by leveraging the local contextual information of related event mentions, generating a more comprehensive global semantic representation of event concepts. At the same time, the knowledge generated by COMET is filtered and enriched using LLM, strengthening the associations between event pairs and knowledge. Finally, the joint event conceptual representation and knowledge-enhanced event representation are used to uncover potential causal relationships between events. The experimental results show that our method outperforms previous state-of-the-art methods on both benchmarks, EventStoryLine and Causal-TimeBank.

Event Causal Identification (ECI) aims to identify fine-grained causal relationships between events from unstructured text. Contrastive learning has shown promise in enhancing ECI by optimizing representation distances between positive and negative samples. However, existing methods often rely on rule-based or random sampling strategies, which may introduce spurious causal positives. Moreover, static negative samples often fail to approximate actual decision boundaries, thus limiting discriminative performance. Therefore, we propose an ECI method enhanced by Dynamic Energy-based Contrastive Learning with multi-stage knowledge Verification (DECLV). Specifically, we integrate multi-source knowledge validation and LLM-driven causal inference to construct a multi-stage knowledge validation mechanism, which generates high-quality contrastive samples and effectively suppresses spurious causal disturbances. Meanwhile, we introduce the Stochastic Gradient Langevin Dynamics (SGLD) method to dynamically generate adversarial negative samples, and employ an energy-based function to model the causal boundary between positive and negative samples. The experimental results show that our method outperforms previous state-of-the-art methods on both benchmarks, EventStoryLine and Causal-TimeBank.

As a promising strategy to achieve generalizable graph learning tasks, graph invariant learning emphasizes identifying invariant subgraphs for stable predictions on biased unknown distribution by selecting the important edges/nodes based on their contributions to the predictive tasks (i.e., subgraph predictivity). However, the existing approaches solely relying on subgraph predictivity face a challenge: the learned invariant subgraph often contains numerous spurious nodes and shows poor connectivity, undermining the generalization power of Graph Neural Networks (GNNs). To tackle this issue, we propose a summary graph-induced Invariant Learning (SIL) model that innovatively adopts a summary graph to leverage both the subgraph connectivity and predictivity for learning strong connected and accurate invariant subgraphs. Specifically, SIL first learns a summary graph containing multiple strongly connected supernodes while maintaining structure consistency with the original graph. Second, the learned summary graph is disentangled into an invariant supernode and spurious counterparts to eliminate the interference of highly predictive edges and nodes. Finally, SIL identifies a potential invariant subgraph from the invariant supernode to accomplish generalization tasks. Additionally, we provide a theoretical analysis of the summary graph learning mechanism, guaranteeing that the learned summary graph is consistent with the original graph. Experimental results validate the effectiveness of the SIL model.

Event Argument Extraction (EAE) aims to extract arguments for specified events from a text.Previous research has mainly focused on addressing long-distance dependencies of arguments, modeling co-occurrence relationships between roles and events, but overlooking potential inductive biases: (i) semantic differences among arguments of the same type and (ii) large margin separation between arguments of the different types.Inspired by prototype networks, we introduce a new model named HMPEAE, which takes the two inductive biases above as targets to locate prototypes and guide the model to learn argument representations based on these prototypes.Specifically, we set multiple prototypes to represent each role to capture intra-class differences.Simultaneously, we use hypersphere as the output space for prototypes, defining large margin separation between prototypes to encourage the model to learn significant differences between different types of arguments effectively.We solve the "argument-prototype" assignment as an optimal transport problem to optimize the argument representation and minimize the absolute distance between arguments and prototypes to achieve compactness within sub-clusters.Experimental results on the RAMS and WikiEvents datasets show that HMPEAE achieves state-of-the-art performances.

Commonsense question answering (CQA) generally means that the machine uses its mastered commonsense to answer questions without relevant background material, which is a challenging task in natural language processing. Existing methods focus on retrieving relevant subgraphs from knowledge graphs based on key entities and designing complex graph neural networks to perform reasoning over the subgraphs. However, they have the following problems: i) the nested entities in key entities lead to the introduction of irrelevant knowledge; ii) the QA context is not well integrated with the subgraphs; and iii) insufficient context knowledge hinders subgraph nodes understanding. In this paper, we present a heterogeneous-graph reasoning with context paraphrase method (HCP), which introduces the paraphrase knowledge from the dictionary into key entity recognition and subgraphs construction, and effectively fuses QA context and subgraphs during the encoding phase of the pre-trained language model (PTLM). Specifically, HCP filters the nested entities through the dictionary's vocabulary and constructs the Heterogeneous Path-Paraphrase (HPP) graph by connecting the paraphrase descriptions11The paraphrase descriptions are English explanations of words or phrases in WordNet and Wiktionary.with the key entity nodes in the subgraphs. Then, by constructing the visible matrices in the PTLM encoding phase, we fuse the QA context representation into the HPP graph. Finally, to get the answer, we perform reasoning on the HPP graph by Mask Self-Attention. Experimental results on CommonsenseQA and OpenBookQA show that fusing QA context with HPP graph in the encoding stage and enhancing the HPP graph representation by using context paraphrase can improve the machine's commonsense reasoning ability.

Recently, knowledge graphs (KGs) have won noteworthy success in commonsense question answering.Existing methods retrieve relevant subgraphs in the KGs through key entities and reason about the answer with language models (LMs) and graph neural networks.However, they ignore (i) optimizing the knowledge representation and structure of subgraphs and (ii) deeply fusing heterogeneous QA context with subgraphs.In this paper, we propose a dynamic heterogeneous-graph reasoning method with LMs and knowledge representation learning (DHLK), which constructs a heterogeneous knowledge graph (HKG) based on multiple knowledge sources and optimizes the structure and knowledge representation of the HKG using a two-stage pruning strategy and knowledge representation learning (KRL).It then performs joint reasoning by LMs and Relation Mask Self-Attention (RMSA).Specifically, DHLK filters key entities based on the dictionary vocabulary to achieve the first-stage pruning while incorporating the paraphrases in the dictionary into the subgraph to construct the HKG.Then, DHLK encodes and fuses the QA context and HKG using LM, and dynamically removes irrelevant KG entities based on the attention weights of LM for the second-stage pruning.Finally, DHLK introduces KRL to optimize the knowledge representation and perform answer reasoning on the HKG by RMSA.We evaluate DHLK at CommonsenseQA and OpenBookQA, and show its improvement on existing LM and LM+KG methods.

Learning a causal structure from observational data is crucial for data scientists. Recent advances in causal structure learning (CSL) have focused on local-to-global learning, since the local-to-global CSL can be scaled to high-dimensional data. The local-to-global CSL algorithms first learn the local skeletons, then construct the global skeleton, and finally orient edges. In practice, the performance of local-to-global CSL mainly depends on the accuracy of the global skeleton. However, in many real-world settings, owing to inevitable data quality issues (e.g. noise and small sample), existing local-to-global CSL methods often yield many asymmetric edges (e.g., given anasymmetric edge containing variables A and B, the learned skeleton of A contains B, but the learned skeleton of B does not contain A), which make it difficult to construct a high quality global skeleton. To tackle this problem, this paper proposes a Bootstrap sampling based Causal Structure Learning (BCSL) algorithm. The novel contribution of BCSL is that it proposes an integrated global skeleton learning strategy that can construct more accurate global skeletons. Specifically, this strategy first utilizes the Bootstrap method to generate multiple sub-datasets, then learns the local skeleton of variables on each asymmetric edge on those sub-datasets, and finally designs a novel scoring function to estimate the learning results on all sub-datasets for correcting the asymmetric edge. Extensive experiments on both benchmark and real datasets verify the effectiveness of the proposed method.