Tourist Guidance Robot Based on HyperCLOVA

TL;DR

This paper presents a hybrid dialogue system using HyperCLOVA for response generation and information processing, integrated with a custom speech recognition system, achieving high competition rankings.

Contribution

The novel integration of HyperCLOVA with rule-based and generation-based components for Japanese dialogue systems.

Findings

Ranked second in the preliminary round of the competition

Effective use of HyperCLOVA for multiple dialog tasks

Successful combination of rule-based and generation-based methods

Abstract

This paper describes our system submitted to Dialogue Robot Competition 2022. Our proposed system is a combined model of rule-based and generation-based dialog systems. The system utilizes HyperCLOVA, a Japanese foundation model, not only to generate responses but also summarization, search information, etc. We also used our original speech recognition system, which was fine-tuned for this dialog task. As a result, our system ranked second in the preliminary round and moved on to the finals.