[2603.02267] Boosting Meta-Learning for Few-Shot Text Classification via Label-guided Distance Scaling

Computer Science > Machine Learning arXiv:2603.02267 (cs) [Submitted on 28 Feb 2026] Title:Boosting Meta-Learning for Few-Shot Text Classification via Label-guided Distance Scaling Authors:Yunlong Gao, Xinyue Liu, Yingbo Wang, Linlin Zong, Bo Xu View a PDF of the paper titled Boosting Meta-Learning for Few-Shot Text Classification via Label-guided Distance Scaling, by Yunlong Gao and 4 other authors View PDF HTML (experimental) Abstract:Few-shot text classification aims to recognize unseen classes with limited labeled text samples. Existing approaches focus on boosting meta-learners by developing complex algorithms in the training stage. However, the labeled samples are randomly selected during the testing stage, so they may not provide effective supervision signals, leading to misclassification. To address this issue, we propose a \textbf{L}abel-guided \textbf{D}istance \textbf{S}caling (LDS) strategy. The core of our method is exploiting label semantics as supervision signals in both the training and testing stages. Specifically, in the training stage, we design a label-guided loss to inject label semantic information, pulling closer the sample representations and corresponding label representations. In the testing stage, we propose a Label-guided Scaler which scales sample representations with label semantics to provide additional supervision signals. Thus, even if labeled sample representations are far from class centers, our Label-guided Scaler pulls them closer to th...