Ambiguous Learning from Retrieval: Towards Zero-shot Semantic Parsing

Current neural semantic parsers mostly take supervised approaches, which require a considerable amount of expensive training data. As a result, minimizing supervision requirements has been one of the key challenges in semantic parsing. In this paper, we propose a Retrieval as Ambiguous Supervision framework, which can effectively collect high-coverage ambiguous supervisions (i.e., the parse candidates of an utterance) via a pre-trained language modelsbased retrieval system. Then, by assuming candidates will contain the correct ones, the zeroshot task can be converted into an ambiguously supervised task. To improve the precision and coverage of such ambiguous supervision, we propose a confidence-driven self-training algorithm, in which a semantic parser is learned and exploited to disambiguate candidates iteratively. Experimental results show that our approach significantly outperforms the state-of-the-art zero-shot semantic parsing methods.