Unsupervised Candidate Answer Extraction through Differentiable Masker-Reconstructor Model
Submission Type: Regular Long Paper
Submission Track: Information Extraction
Submission Track 2: Question Answering
Keywords: Candidate Answer Extraction, Self-consitency Learning, Unsupervised Learning, Masker-Reconstructor Model
TL;DR: Unsupervised Differentiable Masker-Reconstructor(DMR) model achieves competitive performance on candidate answer extraction task.
Abstract: Question generation is a widely used data augmentation approach with extensive applications, and extracting qualified candidate answers from context passages is a critical step for most question generation systems. However, existing methods for candidate answer extraction are reliant on linguistic rules or annotated data that face the partial annotation issue and challenges in generalization. To overcome these limitations, we propose a novel unsupervised candidate answer extraction approach that leverages the inherent structure of context passages through a Differentiable Masker-Reconstructor (DMR) Model with the enforcement of self-consistency for picking up salient information tokens. We curated two datasets with exhaustively-annotated answers and benchmark a comprehensive set of supervised and unsupervised candidate answer extraction methods. We demonstrate the effectiveness of the DMR model by showing its performance is superior among unsupervised methods and comparable to supervised methods.
Submission Number: 4050
Loading