PAST-TIDE: Prototype-Anchored Statement Tuning with Topic-Invariant Normalization for Stance Detection
2026-07-06 • Computation and Language
Computation and LanguageMachine Learning
AI summaryⓘ
The authors present PAST-TIDE, a system designed to detect stances in Arabic text related to the Nakba topic. Instead of adding a new classification layer, they use a clever way to turn labels into masked words, letting a language model predict them directly. They also improve training with special contrastive learning and a normalization technique that adjusts for different topics. Their approach performs well on official tests, showing that small tweaks to existing models can work effectively when data is limited.
stance detectionmasked language modelingverbalizerprototypical contrastive learninglayer normalizationArabic NLPlow-resource settingsNakbaNLPmacro-F1 scoreshared task
Authors
Md. Shakhoyat Rahman Shujon, MD Jahid Hasan Jim, Md. Milon Islam, Md Rezwanul Haque, Fakhri Karray
Abstract
We introduce PAST-TIDE, our stance detection system addressing both subtasks of the StanceNakba Shared Task at NakbaNLP@LREC-COLING 2026. The main idea is statement tuning. We redefine stance as cloze-style masked language modeling (MLM), letting a verbalizer map label words to stance categories through the pre-trained MLM head rather than appending a randomly initialized classification head. We complement this with prototypical contrastive learning, which uses learnable class prototypes for batch-size independent contrastive training, and topic-conditional layer normalization for cross-topic Arabic stance detection. PAST-TIDE achieves macro-F1 scores of 0.75 for Subtask A and 0.74 for Subtask B on the official leaderboard, indicating that minimal architectural additions to a pre-trained model can remain competitive in low-resource settings.