Online Methods for Multi-Domain Learning and Adaptation

NLP tasks are often domain specific, yet systems can learn behaviors across multiple domains. We develop a new multi-domain online learning framework based on parameter combination from multiple classifiers. Our algorithms draw from multi-task learning and domain adaptation to adapt multiple source domain classifiers to a new target domain, learn across multiple similar domains, and learn across a large number of disparate domains. We evaluate our algorithms on two popular NLP domain adaptation tasks: sentiment classification and spam filtering.

• Type: Conference
• Year: 2008
• Where: EMNLP
• Authors: Mark Dredze, Koby Crammer
• Keywords: Domain Adaptation, EMNLP 2008, Multiple Source Domain, Natural Language Processing, NLP Domain Adaptation