Neural-based Context Representation Learning for Dialog Act Classification

TL;DR

This paper investigates neural methods for dialog act classification, focusing on context representation learning using RNNs and attention mechanisms, demonstrating improved performance across datasets with context-aware models.

Contribution

It introduces and compares various neural architectures combining RNNs and attention mechanisms for context representation in dialog act classification.

Findings

Context-aware models outperform non-contextual ones.

The effectiveness of attention mechanisms varies with dataset characteristics.

Certain attention architectures are more suitable depending on dataset nature.

Abstract

We explore context representation learning methods in neural-based models for dialog act classification. We propose and compare extensively different methods which combine recurrent neural network architectures and attention mechanisms (AMs) at different context levels. Our experimental results on two benchmark datasets show consistent improvements compared to the models without contextual information and reveal that the most suitable AM in the architecture depends on the nature of the dataset.