Non-linear models recently receive a lot of attention as people are startingto discover the power of statistical and embedding features. However,tree-based models are seldom studied in the context of structured learningdespite their recent success on various classification and ranking tasks. Inthis paper, we propose S-MART, a tree-based structured learning framework basedon multiple additive regression trees. S-MART is especially suitable forhandling tasks with dense features, and can be used to learn many differentstructures under various loss functions. We apply S-MART to the task of tweet entity linking --- a core component oftweet information extraction, which aims to identify and link name mentions toentities in a knowledge base. A novel inference algorithm is proposed to handlethe special structure of the task. The experimental results show that S-MARTsignificantly outperforms state-of-the-art tweet entity linking systems.
展开▼
