Recurrent Neural Network Grammars
This addresses the need for improved parsing and language modeling in natural language processing, representing a significant advance rather than an incremental improvement.
The paper tackled the problem of modeling sentences with explicit phrase structure by introducing recurrent neural network grammars, achieving better parsing in English than any previous supervised generative model and superior language modeling compared to state-of-the-art sequential RNNs in English and Chinese.
We introduce recurrent neural network grammars, probabilistic models of sentences with explicit phrase structure. We explain efficient inference procedures that allow application to both parsing and language modeling. Experiments show that they provide better parsing in English than any single previously published supervised generative model and better language modeling than state-of-the-art sequential RNNs in English and Chinese.