arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2103.01273cs.CL
收藏

单语言、多语言和Zero-shot条件下数据集嵌入的有效性研究

On the Effectiveness of Dataset Embeddings in Mono-lingual,Multi-lingual and Zero-shot Conditions

IT University of Copenhagen(哥本哈根信息技术大学);University of Groningen(格罗宁根大学)

Rob van der Goot, Ahmet Üstün, Barbara Plank

摘要:

最近的补充研究表明,通过将数据源的属性编码到嵌入中来利用数据源上的信息,可以在异构数据源上训练单个模型时提高性能。然而,目前还不清楚这些数据集嵌入在哪些情况下最有效,因为它们用于各种各样的设置,语言和任务。此外,通常假设关于数据源的黄金信息是可用的,并且测试数据来自训练期间看到的分布。在这项工作中,我们比较了数据集嵌入在单语言设置,多语言设置,并与预测的数据源标签在zero-shot设置的效果。我们评估三个形态句法任务:形态标记,词形化,和依存句法分析,并使用104个数据集,66种语言,和两种不同的数据集分组策略。当数据集是同一种语言时,性能提高最高,并且我们知道测试实例来自哪个分布。相比之下,对于数据来自不可见分布的设置,性能提高消失了。

英文摘要:

Recent complementary strands of research have shown that leveraging information on the data source through encoding their properties into embeddings can lead to performance increase when training a single model on heterogeneous data sources. However, it remains unclear in which situations these dataset embeddings are most effective, because they are used in a large variety of settings, languages and tasks. Furthermore, it is usually assumed that gold information on the data source is available, and that the test data is from a distribution seen during training. In this work, we compare the effect of dataset embeddings in mono-lingual settings, multi-lingual settings, and with predicted data source label in a zero-shot setting. We evaluate on three morphosyntactic tasks: morphological tagging, lemmatization, and dependency parsing, and use 104 datasets, 66 languages, and two different dataset grouping strategies. Performance increases are highest when the datasets are of the same language, and we know from which distribution the test-instance is drawn. In contrast, for setups where the data is from an unseen distribution, performance increase vanishes.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑