Efficient Second-Order TreeCRF for Neural Dependency Parsing
This work addresses the problem of enhancing dependency parsing accuracy for NLP researchers and practitioners by incorporating structural learning techniques, though it is incremental as it builds on existing biaffine parser methods.
The paper tackles the challenge of integrating second-order TreeCRF into neural dependency parsing by developing efficient GPU-optimized algorithms, resulting in improved parsing performance over the state-of-the-art biaffine parser, especially on partially annotated data across 27 datasets from 13 languages.
In the deep learning (DL) era, parsing models are extremely simplified with little hurt on performance, thanks to the remarkable capability of multi-layer BiLSTMs in context representation. As the most popular graph-based dependency parser due to its high efficiency and performance, the biaffine parser directly scores single dependencies under the arc-factorization assumption, and adopts a very simple local token-wise cross-entropy training loss. This paper for the first time presents a second-order TreeCRF extension to the biaffine parser. For a long time, the complexity and inefficiency of the inside-outside algorithm hinder the popularity of TreeCRF. To address this issue, we propose an effective way to batchify the inside and Viterbi algorithms for direct large matrix operation on GPUs, and to avoid the complex outside algorithm via efficient back-propagation. Experiments and analysis on 27 datasets from 13 languages clearly show that techniques developed before the DL era, such as structural learning (global TreeCRF loss) and high-order modeling are still useful, and can further boost parsing performance over the state-of-the-art biaffine parser, especially for partially annotated training data. We release our code at https://github.com/yzhangcs/crfpar.