Fine-grained Contrastive Learning for Definition Generation
This work addresses a specific bottleneck in definition generation for natural language processing, offering an incremental improvement over existing methods.
The paper tackles the problem of generating under-specific definitions in definition generation by proposing a novel contrastive learning method to capture detailed semantic representations, resulting in more specific and high-quality definitions on three benchmarks.
Recently, pre-trained transformer-based models have achieved great success in the task of definition generation (DG). However, previous encoder-decoder models lack effective representation learning to contain full semantic components of the given word, which leads to generating under-specific definitions. To address this problem, we propose a novel contrastive learning method, encouraging the model to capture more detailed semantic representations from the definition sequence encoding. According to both automatic and manual evaluation, the experimental results on three mainstream benchmarks demonstrate that the proposed method could generate more specific and high-quality definitions compared with several state-of-the-art models.