LogSigma at SemEval-2026 Task 3: Uncertainty-Weighted Multitask Learning for Dimensional Aspect-Based Sentiment Analysis
This work addresses the problem of language-dependent prediction difficulty in sentiment analysis for researchers and practitioners, representing an incremental improvement through uncertainty-weighted multitask learning.
The paper tackled the challenge of predicting continuous Valence and Arousal scores in Dimensional Aspect-Based Sentiment Analysis by using learned homoscedastic uncertainty to automatically balance regression objectives, achieving first place on five datasets across both tracks with variance weights ranging from 0.66x for German to 2.18x for English.
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.