Agree to Disagree: Improving Disagreement Detection with Dual GRUs

TL;DR

This paper introduces a dual GRU-based model for disagreement detection in online discussions, eliminating the need for hand-crafted features and achieving state-of-the-art results on multiple datasets.

Contribution

The paper proposes a Siamese-inspired dual GRU architecture that encodes discussions without relying on manual features, improving disagreement detection performance.

Findings

Achieved 0.804 F1 score on ABCD dataset.

Model trained on ABCD performs well on smaller datasets.

Fusion of lexical and embedding features enhances accuracy.

Abstract

This paper presents models for detecting agreement/disagreement in online discussions. In this work we show that by using a Siamese inspired architecture to encode the discussions, we no longer need to rely on hand-crafted features to exploit the meta thread structure. We evaluate our model on existing online discussion corpora - ABCD, IAC and AWTP. Experimental results on ABCD dataset show that by fusing lexical and word embedding features, our model achieves the state of the art performance of 0.804 average F1 score. We also show that the model trained on ABCD dataset performs competitively on relatively smaller annotated datasets (IAC and AWTP).