This paper makes a focused contribution to supervised aspect extraction. Itshows that if the system has performed aspect extraction from many past domainsand retained their results as knowledge, Conditional Random Fields (CRF) canleverage this knowledge in a lifelong learning manner to extract in a newdomain markedly better than the traditional CRF without using this priorknowledge. The key innovation is that even after CRF training, the model canstill improve its extraction with experiences in its applications.
展开▼