EMNLP 2016 Papers

We propose a neural-network based model for coordination boundary prediction. The network is designed to incorporate two signals: the similarity between conjuncts and the observation that replacing the whole coordination phrase with a conjunct tends to produce a coherent sentences. The modeling makes use of several LSTM networks. The model is trained solely on conjunction annotations in a Treebank, without using external resources. We show improvements on predicting coordination boundaries on the PTB compared to two state-of-the-art parsers; as well as improvement over previous coordination boundary prediction systems on the Genia corpus.

Tree transducers that model expressive linguistic phenomena often require word-alignments and a heuristic rule extractor to induce their grammars. However, when the corpus of tree/string pairs is small compared to the size of the vocabulary or the complexity of the grammar, word-alignments are unreliable. We propose a general rule extraction algorithm that uses cost functions over tree fragments, and formulate the extraction as a cost minimization problem. As a by-product, we are able to introduce back-off states at which some cost functions generate right-hand-sides of previously unseen left-hand-sides, thus creating transducer rules “on-the-fly”. We test the generalization power of our induced tree transducers on a QA task over a large Knowledge Base, obtaining a reasonable syntactic accuracy and effectively overcoming the typical lack of rule coverage.

Parsing accuracy using efficient greedy transition systems has improved dramatically in recent years thanks to neural networks. Despite striking results in dependency parsing, however, neural models have not surpassed state-of-the-art approaches in constituency parsing. To remedy this, we introduce a new shift-reduce system whose stack contains merely sentence spans, represented by a bare minimum of LSTM features. We also design the first provably optimal dynamic oracle for constituency parsing, which runs in amortized O(1) time, compared to O(n^3) oracles for standard dependency parsing. Training with this oracle, we achieve the best F1 scores on both English and French of any parser that does not use reranking or external data.