Improved intent classification based on context information using a windows-based approach

TL;DR

This paper introduces a window-based CNN approach utilizing BERT embeddings to incorporate dialogue context for improved intent classification in conversational systems, demonstrating significant performance gains on a real-world Portuguese dataset.

Contribution

It proposes a novel window-based method that effectively integrates contextual dialogue information into intent classification using CNNs and BERT, outperforming previous approaches that ignore context.

Findings

Significant accuracy improvements over baseline models.

Effective use of dialogue history from previous utterances.

Validated on a real-world Portuguese dataset.

Abstract

Conversational systems have a Natural Language Understanding (NLU) module. In this module, there is a task known as an intent classification that aims at identifying what a user is attempting to achieve from an utterance. Previous works use only the current utterance to predict the intent of a given query and they do not consider the role of the context (one or a few previous utterances) in the dialog flow for this task. In this work, we propose several approaches to investigate the role of contextual information for the intent classification task. Each approach is used to carry out a concatenation between the dialogue history and the current utterance. Our intent classification method is based on a convolutional neural network that obtains effective vector representations from BERT to perform accurate intent classification using an approach window-based. Our experiments were carried out on a real-world Brazilian Portuguese corpus with dialog flows provided by Wavy global company. Our results achieved substantial improvements over the baseline, isolated utterances (without context), in three approaches using the user's utterance and system's response from previous messages as dialogue context.