Easy-First Dependency Parsing with Hierarchical Tree LSTMs

By Eliyahu Kiperwasser, Yoav Goldberg. 2016

67 citations recorded by the source corpus.

We suggest a compositional vector representation of parse trees that relies on a recursive combination of recurrent-neural network encoders. To demonstrate its effectiveness, we use the representation as the backbone of a greedy, bottom-up dependency parser, achieving very strong accuracies for English and Chinese, without relying on external word embeddings. The parser’s implementation is available for download at the first author’s webpage.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Natural Language Processing, Model Architectures

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Natural Language Processing

In Model Architectures

Citations in the corpus

Recurrent Neural Network Grammars (2016, background): “Syntactically structured neural architectures have been explored in a number of applications, including discriminative parsing (Socher et al., 2013a; Kiperwasser and Goldberg, 2016), sentiment analysis (Tai et al., 2015; Socher et al., 2013b), and sentence representation (Socher et al., 2011;…”
A Joint Many-Task Model: Growing a Neural Network for Multiple NLP Tasks (2016, background): “We call the second regularization term δ (cid:107) θ e − θ e 2 a successive regularization term.”
A Fast Unified Model for Parsing and Sentence Understanding (2016, background): “…that use the core operations and data structures from transition-based parsing, of which shift-reduce parsing is a variant (Henderson, 2004; Emami and Jelinek, 2005; Titov and Henderson, 2010; Chen and Manning, 2014; Buys and Blunsom, 2015; Dyer et al., 2015; Kiperwasser and Goldberg, 2016).”
Junction Tree Variational Autoencoder for Molecular Graph Generation (2018, methodology): “The first two evaluations follow previously proposed tasks (Kusner et al., 2017). We also introduce a third task — constrained molecule optimization. • Molecule reconstruction and validity We test the VAE models on the task of reconstructing input molecules from their latent representations, and dec”
A Neural Probabilistic Structured-Prediction Method for Transition-Based Natural Language Processing (2017, methodology): “The method of Kudo and Matsumoto (2001) gives the highest result in CoNLL 2000, which adopts the SVM training and dynamic programming decoding.”
Learning What’s Easy: Fully Differentiable Neural Easy-First Taggers (2017, methodology): “Usually, gradients are not backpropagated over this combinatorial latent space (Kiperwasser and Goldberg, 2016), or a separate model is used to determine the easiest next move (Clark and Manning, 2016).”
Tree-structured decoding with doubly-recurrent neural networks (2016, background): “[1, 3]) In case (a), the network aggregates the node information in a manner that is coherent with a given tree structure.”