On Graph-based Reentrancy-free Semantic Parsing

Abstract We propose a novel graph-based approach for semantic parsing that resolves two problems observed in the literature: (1) seq2seq models fail on compositional generalization tasks; (2) previous work using phrase structure parsers cannot cover all parses treebanks. prove both MAP inference and latent tag anchoring (required weakly-supervised learning) are NP-hard problems. optimization algorithms based constraint smoothing conditional gradient to approximately solve these Experimentally, our delivers state-of-the-art results GeoQuery, Scan, Clevr, i.i.d. splits test generalization.