MCPG: A Flexible Multi-Level Controllable Framework for Unsupervised Paraphrase Generation

We present MCPG: a simple and effective approach for controllable unsupervised paraphrase generation, which is also flexible to adapt to specific domains without extra training. MCPG is controllable in different levels: local lexicons, global semantics, and universal styles. The unsupervised paradigm of MCPG combines factual keywords and diversified semantic embeddings as local lexical and global semantic constraints. The semantic embeddings are diversified by standard dropout, which is exploited for the first time to increase inference diversity by us. Moreover, MCPG is qualified with good domain adaptability by adding a transfer vector as a universal style constraint, which is refined from the exemplars retrieved from the corpus of the target domain in a training-free way. Extensive experiments show that MCPG outperforms state-of-the-art unsupervised baselines by a margin. Meanwhile, our domain-adapted MCPG also achieves competitive performance with strong supervised baselines even without training.