MS MARCO Web Search
收藏资源简介:
MS MARCO Web Search是由微软创建的大型信息丰富的网络数据集,包含数百万真实的点击查询-文档标签。该数据集模拟了真实世界的网络文档和查询分布,提供了丰富的信息用于各种下游任务,并鼓励在多个领域进行研究,如通用端到端神经索引模型、通用嵌入模型和下一代大型语言模型信息访问系统。作为首个满足大数据、真实和丰富数据要求的数据集,MS MARCO Web Search为未来的AI和系统研究提供了关键的数据基础。
MS MARCO Web Search is a large-scale, information-rich web dataset developed by Microsoft, which contains millions of real click-derived query-document labels. This dataset replicates the real-world distribution of web documents and user queries, providing rich resources for a wide range of downstream tasks and facilitating research across multiple domains, including general-purpose end-to-end neural indexing models, general embedding models, and next-generation large language model-driven information access systems. As the first dataset that meets the criteria of large-scale, authentic, and rich data, MS MARCO Web Search serves as a critical data foundation for future AI and system-related research.

- 1MS MARCO Web Search: a Large-scale Information-rich Web Dataset with Millions of Real Click Labels微软 · 2024年



