Improving Lexical Difficulty Prediction with Context-Aligned Contrastive Learning and Ridge Ensembling
For researchers in language learning and readability assessment, this work offers a method to better predict word difficulty across different first-language backgrounds, though the gains are incremental over existing regression-only approaches.
The paper proposes Context-Aligned Contrastive Regression, combining Ridge regression ensemble with contrastive learning objectives, to improve lexical difficulty prediction across L1 backgrounds. Experiments on three datasets show improved cross-lingual alignment and ordinal structure capture, with more stable performance across difficulty levels.
Lexical difficulty prediction is a fundamental problem in language learning and readability assessment, requiring models to estimate word difficulty across different first-language (L1) backgrounds. However, existing approaches rely on regression-only training with scalar supervision, which does not explicitly structure the representation space, limiting their ability to capture cross-lingual alignment and ordinal difficulty. To mitigate these issues, we propose Context-Aligned Contrastive Regression, which integrates Ridge regression ensemble with two complementary objectives, i.e., Cross-View Context and Ordinal Soft Contrastive Learning. Experiments on three L1 datasets show that (i) contrastive objectives improve cross-lingual representation alignment while preserving language-specific nuances, (ii) the learned representations capture the ordinal structure of lexical difficulty, and (iii) the ensemble effectively mitigates systematic biases of individual models, leading to more stable performance across difficulty levels.