Learn More
This paper describes our participation in Knowledge Base Population track at TAC2010. In the entity-linking task, we combined machine learning-based methods and rule-based methods to improve the linking results. In the slot filling task, a supervised machine learning method based on CRF model and a rule pattern method were used to select proper answers for(More)
This paper describes BUPT (pris) participation in entity linking task and slot filling task. The system adopts a two-stage strategy in entity linking task and slot filling task. In the first stage, the system carries out a basic topic relevance retrieval to get top k documents for each query. In the second stage, cross-document coreference resolution is(More)
This paper presents the system adopted for the Faceted Blog Distillation task by PRIS team. The PRIS system is submitted by Pattern Recognition and Intelligent System Lab at Beijing University of Posts and Telecommunications. And a two-stage strategy is involved for this task. First, an adaptable Voting Model is carried out for blog distillation. Then,(More)
With the increasing of blog users, the traditional blog search can no longer meet their demands. More work should be done to accommodate the need of finding good blogs to read, besides the topic-relevant blogs. This paper focuses on the problem of an in-depth faceted blog distillation for addressing the quality aspect of the retrieval blogs. We propose a(More)
  • 1