DyREx: Dynamic Query Representation for Extractive Question Answering

TL;DR

DyREx introduces a dynamic query vector computation method for extractive question answering, leveraging attention mechanisms to incorporate input context, leading to improved performance over traditional static query approaches.

Contribution

The paper proposes DyREx, a novel method that dynamically computes query vectors using attention, enhancing extractive QA models beyond static query vector techniques.

Findings

Consistently outperforms standard static query models

Improves answer span prediction accuracy

Demonstrates effectiveness across datasets

Abstract

Extractive question answering (ExQA) is an essential task for Natural Language Processing. The dominant approach to ExQA is one that represents the input sequence tokens (question and passage) with a pre-trained transformer, then uses two learned query vectors to compute distributions over the start and end answer span positions. These query vectors lack the context of the inputs, which can be a bottleneck for the model performance. To address this problem, we propose \textit{DyREx}, a generalization of the \textit{vanilla} approach where we dynamically compute query vectors given the input, using an attention mechanism through transformer layers. Empirical observations demonstrate that our approach consistently improves the performance over the standard one. The code and accompanying files for running the experiments are available at \url{https://github.com/urchade/DyReX}.