NCL-BU 在 SemEval-2026 任务 3: 基于 XLM-RoBERTa 的多语言维度情感回归微调
NCL-BU at SemEval-2026 Task 3: Fine-tuning XLM-RoBERTa for Multilingual Dimensional Sentiment Regression
AI总结：
本文提出基于 XLM-RoBERTa 的微调方法，用于多语言维度情感回归任务，通过双回归头预测情感值和唤醒度，跨领域和语言进行训练，实验显示任务特定微调优于其他大语言模型。
摘要：
维度方面情感分析（DimABSA）将传统情感分析从分类极性标签扩展到连续的估值-唤醒（VA）回归。本文描述了为 Track A，Subtask 1（维度方面情感回归）开发的系统，旨在为文本中的每个方面预测 [1, 9] 范围内的实值 VA 分数。采用基于 XLM-RoBERTa-base 的微调方法，使用双回归头进行估值和唤醒度预测，输出经过 sigmoid 缩放。分别针对每种语言-领域对（英语和中文在餐厅、笔记本电脑和金融领域）训练模型，并将训练和开发集合并用于最终测试预测。在开发实验中，微调方法在少样本提示设置下与几种大语言模型进行比较，证明任务特定微调在所有评估数据集中优于这些 LLM 基方法。
英文摘要：
Dimensional Aspect-Based Sentiment Analysis (DimABSA) extends traditional ABSA from categorical polarity labels to continuous valence-arousal (VA) regression. This paper describes a system developed for Track A, Subtask 1 (Dimensional Aspect Sentiment Regression), aiming to predict real-valued VA scores in the [1, 9] range for each given aspect in a text. A fine-tuning approach based on XLM-RoBERTa-base is adopted, with dual regression heads with sigmoid-scaled outputs for valence and arousal prediction. Separate models are trained for each language-domain pair (English and Chinese across restaurant, laptop, and finance domains), and training and development sets are merged for final test predictions. In development experiments, the fine-tuning approach is compared against several large language models under a few-shot prompting setting, demonstrating that task-specific fine-tuning outperforms these LLM-based methods across all evaluation datasets.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
