arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1809.02256cs.CL
收藏

混合专家模型下的多源领域自适应

Multi-Source Domain Adaptation with Mixture of Experts

Computer Science and Artificial Intelligence Laboratory(计算机科学与人工智能实验室);Massachusetts Institute of Technology(麻省理工学院)

Jiang Guo, Darsh J Shah, Regina Barzilay

摘要:

我们提出了一个混合的专家方法,从多个来源的无监督领域自适应。关键思想是显式地捕获目标示例和不同源域之间的关系。这种关系由点到集度量表示,决定了如何组合在各个领域训练的预测器。该指标是使用元训练以无监督的方式学习的。情感分析和词性标注的实验结果表明,我们的方法始终优于多个基线,可以鲁棒地处理负迁移。

英文摘要:

We propose a mixture-of-experts approach for unsupervised domain adaptation from multiple sources. The key idea is to explicitly capture the relationship between a target example and different source domains. This relationship, expressed by a point-to-set metric, determines how to combine predictors trained on various domains. The metric is learned in an unsupervised fashion using meta-training. Experimental results on sentiment analysis and part-of-speech tagging demonstrate that our approach consistently outperforms multiple baselines and can robustly handle negative transfer.

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。

↑