提出新损失函数，让语音评估模型更公平地对待不同水平学习者。
Mitigating Data Imbalance in Automated Speaking Assessment
- 设计平衡逻辑变化损失，调整模型预测以增强少数类表征。
- 在ICNALE数据集上提升分类准确率与评估公平性。
- 无需改数据即可改善模型偏差，适合多语言教育场景使用。
自动口语评估（ASA）在衡量第二语言学习者能力方面至关重要，但常因类别不平衡导致预测偏倚。为此，本文提出一种新型训练目标——平衡逻辑变化（BLV）损失，通过扰动模型预测来改进少数类的特征表示，无需修改数据集。在ICNALE基准数据集上的实验表明，将BLV损失融入经典的文本基模型（BERT）后，显著提升了分类准确率与评估公平性，使自动化语音评估对多元学习者更具鲁棒性。
原文摘要 · Abstract (English)
Automated Speaking Assessment (ASA) plays a crucial role in evaluating second-language (L2) learners proficiency. However, ASA models often suffer from class imbalance, leading to biased predictions. To address this, we introduce a novel objective for training ASA models, dubbed the Balancing Logit Variation (BLV) loss, which perturbs model predictions to improve feature representation for minority classes without modifying the dataset. Evaluations on the ICNALE benchmark dataset show that integrating the BLV loss into a celebrated text-based (BERT) model significantly enhances classification accuracy and fairness, making automated speech evaluation more robust for diverse learners.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。