GADePo: Graph-Assisted Declarative Pooling Transformers for Document-Level Relation Extraction
This addresses the problem of information aggregation in document-level relation extraction for NLP researchers, offering a more flexible and customizable method, though it is incremental as it builds on existing Transformer architectures.
The paper tackles document-level relation extraction by replacing hand-coded pooling methods with a graph-assisted declarative pooling approach using Transformers, resulting in consistently better performance across diverse datasets compared to existing methods.
Document-level relation extraction typically relies on text-based encoders and hand-coded pooling heuristics to aggregate information learned by the encoder. In this paper, we leverage the intrinsic graph processing capabilities of the Transformer model and propose replacing hand-coded pooling methods with new tokens in the input, which are designed to aggregate information via explicit graph relations in the computation of attention weights. We introduce a joint text-graph Transformer model and a graph-assisted declarative pooling (GADePo) specification of the input, which provides explicit and high-level instructions for information aggregation. GADePo allows the pooling process to be guided by domain-specific knowledge or desired outcomes but still learned by the Transformer, leading to more flexible and customisable pooling strategies. We evaluate our method across diverse datasets and models and show that our approach yields promising results that are consistently better than those achieved by the hand-coded pooling functions.