Scholarlink

This work focuses on querying incomplete data graphs using a hybrid, query-driven approach. The method combines the certainty of known paths with the flexibility of on-the-fly embedding-based link prediction guided by the query structure. A lazy-prediction strategy is used to improve relevance while limiting unnecessary inference. We evaluate our method on two embedding-based models and datasets, and test predictive strategies with different search space reductions.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Lazy Prediction in Querying Graph Databases

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Jacques Chabin,
  • Cristina D. Aguiar,
  • Mirian Halfeld-Ferrari,
  • Martin A. Musicante,
  • Lingchen Wang

摘要

This work focuses on querying incomplete data graphs using a hybrid, query-driven approach. The method combines the certainty of known paths with the flexibility of on-the-fly embedding-based link prediction guided by the query structure. A lazy-prediction strategy is used to improve relevance while limiting unnecessary inference. We evaluate our method on two embedding-based models and datasets, and test predictive strategies with different search space reductions.