TargetSpecificDomainAdaptation This is the implementation of the paper "Cross-Domain Sentiment Classification with Target Domain Specific Information" published in ACL 2018. Running Example For peroformance of our proposed method on task "0->1", run: CUDA_VISIBLE_DEVICES=0 python3 co_training.py 0 1 Here, "0, 1, 2, 3" denotes the "book", "dvd", "electronic" and "kitchen" domain, respectively.