Learning Translational and Knowledge-based Similarities from Relevance Rankings for Cross-Language Retrieval
Learning Translational and Knowledge-based Similarities from Relevance Rankings for Cross-Language Retrieval
复制标题
DOI：
10.3115/v1/p14-2080
复制
发表时间：
2014-06
期刊：
影响因子：
--
通讯作者：
Shigehiko Schamoni;F. Hieber;Artem Sokolov;S. Riezler
中科院分区：
文献类型：
--
作者：
Shigehiko Schamoni;F. Hieber;Artem Sokolov;S. Riezler
我们提出了一种结合密集知识特征和稀疏词翻译的跨语言检索方法。这两种特征类型都是在两两排序框架中直接从双语文档的相关性排序中学习到的。在维基百科的专利现有技术搜索和跨语言检索的大规模实验中，我们的方法比仅使用密集或稀疏特征的学习排序，以及结合了最先进的机器翻译和检索的非常有竞争力的基线，产生了相当大的改进。
We present an approach to cross-language retrieval that combines dense knowledgebased features and sparse word translations. Both feature types are learned directly from relevance rankings of bilingual documents in a pairwise ranking framework. In large-scale experiments for patent prior art search and cross-lingual retrieval in Wikipedia, our approach yields considerable improvements over learningto-rank with either only dense or only sparse features, and over very competitive baselines that combine state-of-the-art machine translation and retrieval.