LogSigma 在 SemEval-2026 任务 3:用于维度方面基于情感分析的不确定性加权多任务学习
LogSigma at SemEval-2026 Task 3: Uncertainty-Weighted Multitask Learning for Dimensional Aspect-Based Sentiment Analysis
AI总结:
本文提出 LogSigma 系统,针对 SemEval-2026 任务 3 的维度方面情感分析,通过学习同方差不确定性平衡回归目标,结合语言特定编码器和多种子集成,在五个数据集上取得第一名。
摘要:
本文描述了 LogSigma,我们的系统用于 SemEval-2026 任务 3:维度方面基于情感分析(DimABSA)。与传统基于方面的情感分析(ABSA)不同,DimABSA 需要预测在 1-9 分度上的连续效价和唤醒度(VA)分数。核心挑战在于效价和唤醒度在不同语言和领域中的预测难度不同。我们通过学习的同方差不确定性来解决这一问题,其中模型学习任务特定的对数方差参数以在训练过程中自动平衡每个回归目标。结合语言特定编码器和多种子集成,LogSigma 在五个数据集上两个赛道均取得第一名。学习的方差权重在不同语言中变化显著,由于效价-唤醒度难度轮廓不同,从德语的 0.66 倍到英语的 2.18 倍,这表明最优任务平衡是语言依赖的,不能事先确定。
英文摘要:
This paper describes LogSigma, our system for SemEval-2026 Task 3: Dimensional Aspect-Based Sentiment Analysis (DimABSA). Unlike traditional Aspect-Based Sentiment Analysis (ABSA), which predicts discrete sentiment labels, DimABSA requires predicting continuous Valence and Arousal (VA) scores on a 1-9 scale. A central challenge is that Valence and Arousal differ in prediction difficulty across languages and domains. We address this using learned homoscedastic uncertainty, where the model learns task-specific log-variance parameters to automatically balance each regression objective during training. Combined with language-specific encoders and multi-seed ensembling, LogSigma achieves 1st place on five datasets across both tracks. The learned variance weights vary substantially across languages due to differing Valence-Arousal difficulty profiles-from 0.66x for German to 2.18x for English-demonstrating that optimal task balancing is language-dependent and cannot be determined a priori.
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。