Multimodal Dialogue State Tracking By QA Approach with Data Augmentation

TL;DR

This paper introduces a multimodal QA-based approach with data augmentation for dialogue state tracking in video-related conversations, improving performance on AVSD tasks.

Contribution

It presents a novel QA framework for AVSD, incorporating data augmentation techniques to enhance dialogue understanding and state tracking accuracy.

Findings

Significant performance improvements over baseline models.

Effective data augmentation enhances model robustness.

Demonstrates potential of QA approach in multimodal dialogue tasks.

Abstract

Recently, a more challenging state tracking task, Audio-Video Scene-Aware Dialogue (AVSD), is catching an increasing amount of attention among researchers. Different from purely text-based dialogue state tracking, the dialogue in AVSD contains a sequence of question-answer pairs about a video and the final answer to the given question requires additional understanding of the video. This paper interprets the AVSD task from an open-domain Question Answering (QA) point of view and proposes a multimodal open-domain QA system to deal with the problem. The proposed QA system uses common encoder-decoder framework with multimodal fusion and attention. Teacher forcing is applied to train a natural language generator. We also propose a new data augmentation approach specifically under QA assumption. Our experiments show that our model and techniques bring significant improvements over the baseline model on the DSTC7-AVSD dataset and demonstrate the potentials of our data augmentation techniques.