Syntax-Aware Graph-to-Graph Transformer for Semantic Role Labelling
This work addresses semantic role labelling, a key task in natural language processing, by improving accuracy through better integration of syntactic knowledge, though it is incremental as it builds on existing methods.
The paper tackles semantic role labelling by proposing a syntax-aware graph-to-graph Transformer model that encodes syntactic structure into self-attention, achieving state-of-the-art performance on CoNLL 2005 and 2009 datasets in both in-domain and out-of-domain settings.
Recent models have shown that incorporating syntactic knowledge into the semantic role labelling (SRL) task leads to a significant improvement. In this paper, we propose Syntax-aware Graph-to-Graph Transformer (SynG2G-Tr) model, which encodes the syntactic structure using a novel way to input graph relations as embeddings, directly into the self-attention mechanism of Transformer. This approach adds a soft bias towards attention patterns that follow the syntactic structure but also allows the model to use this information to learn alternative patterns. We evaluate our model on both span-based and dependency-based SRL datasets, and outperform previous alternative methods in both in-domain and out-of-domain settings, on CoNLL 2005 and CoNLL 2009 datasets.