音频文本跨模态表示的无监督改进（英文版）

英文标题：Unsupervised Improvement of Audio-Text Cross-Modal Representations中文摘要：本文研究了使用无配对数据进行无监督学习的方法，结合领域特定的有软标签的对比损失方法可以显著提高跨模态音频 - 文本表示学习的效果及其在零样本分类任务中的性能。英文摘要：Recent advances in using language models to obtain cross-modal audio-textrepresentations have overcome the limitations of conventional trainingapproaches that use predefined labels. This has allowed the community to makeprogress in tasks like zero-shot classification, which would otherwise not bepossible. However, learning such representa