Search papers, labs, and topics across Lattice.
The paper introduces PAST-TIDE, a stance detection system designed for the StanceNakba Shared Task, which innovatively redefines stance detection using cloze-style masked language modeling (MLM) and incorporates prototypical contrastive learning with topic-conditional normalization. This approach allows for effective stance classification by mapping label words to stance categories through a pre-trained MLM head, enhancing performance in low-resource settings. PAST-TIDE achieved competitive macro-F1 scores of 0.75 and 0.74 for the respective subtasks, demonstrating that minimal architectural modifications can yield significant results in stance detection tasks.
Stance detection can be revolutionized by leveraging cloze-style MLM and prototypical learning, achieving high performance even in low-resource contexts.
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.
