The task of event graph completion (EGC) aims to predict the missing relations between two events in a heterogeneous event graph (EG) to enhance its applicability in downstream tasks. Given that event graphs inherently contain both semantic and structural information, and that the effective aggregation of these two types of information can mutually reinforce their expressive power, we propose a **M**ulti-**E**xpert **R**outing (MER) framework, which leverages the mixture of experts (MoE) architecture to dynamically select appropriate experts for collaboratively addressing different key aspects of the EGC task. Specifically, our approach incorporates multiple specialized experts, leveraging prompt learning and a graph attention network, to independently model distinct semantic and structural features within EGs. Furthermore, an adaptive expert routing (AER) module is designed to dynamically fuse and select the missing relation predictions from the experts, thereby completing the EGs. Extensive experiments conducted on the EGC-MAVEN dataset demonstrate that our MER framework outperforms the state-of-the-art competitors.

Fine-grained negative emotion classification suffers from subjective annotation noise and semantically ambiguous hard samples, which severely degrades model performance. Most existing methods cannot distinguish valid hard samples from noise and lack strategies to improve hard sample learnability. To address these issues, we propose a Noise Filtering and Difficulty-adaptive Hard Sample Augmentation (NF-DHSA) framework for fine-grained negative emotion classification. It first mines high-risk candidates using a hybrid score of normalized loss and predictive entropy, then applies a multi-view logic-guided filtering mechanism to reliably separate noise from valid hard samples. Further, difficulty-adaptive dynamic augmentation generates moderate-difficulty, label-stable augmented samples, and prototype-guided dual-granularity contrastive learning aligns original-augmented pairs while strengthening global category boundaries. The results reported on three benchmark datasets show that NF-DHSA outperforms the baseline methods overall, with Macro-F1 scores improved by 1.25%, 2.47%, and 3.53% on the VIRUS, NLPCC, and COVID datasets, respectively.

Tokenizers play a critical role in large language model studies. Despite recent advances, existing tokenizers fail to explicitly leverage historical tokenization results when making subsequent token decisions, nor do they selectively utilize such history based on contextual relevance. We propose SPEAK, a tokenizer that integrates spiking neurons to explicitly leverage historical tokenization results. Furthermore, we introduce an entropy-aware reset mechanism that selectively leverages history based on contextual relevance, which is determined by token-level entropy. High-entropy tokens are treated as contextual boundaries, whereas low-entropy tokens between consecutive such boundaries exhibit strong contextual relevance. Accordingly, we induce hard reset at high-entropy tokens to discard irrelevant historical tokenization results, and soft reset at low-entropy tokens to preserve and leverage relevant history. Experiments on 2 language models and 5 datasets spanning 16 languages demonstrate superior cross-lingual adaptability, with competitive performance and efficiency. Our code is publicly available at https://github.com/zju-bmi-lab/SPEAK.

Implicit Discourse Relation Recognition (IDRR) aims at classifying the relation sense between two arguments without an explicit connective. Recently, the ConnPrompt (Xiang et al., 2022) has leveraged the powerful prompt learning for IDRR based on the fusion of multi-prompt decisions from three different yet much similar connective prediction templates. Instead of multi-prompt ensembling, we propose to design auxiliary tasks with enlightened prompt learning for the IDRR task. Although an auxiliary task is not used to directly output final prediction, we argue that during the joint training some of its learned features can be useful to boost the main task. In light of such motivations, we propose a task enlightenment prompt learning model, called TEPrompt, to fuse learned features from three related tasks for IDRR. In particular, the TEPrompt contains three tasks, viz., Discourse Relation Recognition (DRR), Sense Semantics Classification (SSC) and Annotated Connective Prediction (ACP), each with a unique prompt template and an answer space. In the training phase, we jointly train three prompt learning tasks with shared argument representation. In the testing phase, we only take the DRR output with fused features as the final IDRR decision. Experiments with the same conditions have shown that the proposed TEPrompt outperforms the ConnPrompt. This can be attributed to the promoted decision features and language models benefited from joint-training of auxiliary tasks.