DORIC : Domain Robust Fine-Tuning for Open Intent Clustering through Dependency Parsing

TL;DR

This paper introduces DORIC, a method for open intent clustering that uses dependency parsing and fine-tuning to improve cross-domain intent recognition without in-domain training data.

Contribution

It proposes a novel approach combining dependency parsing and multi-domain fine-tuning to enhance zero-shot intent clustering and explainability in dialogue systems.

Findings

Achieved 3rd place in precision score at DSTC11-Track2

Outperformed baseline models in accuracy and NMI scores

Effective in cross-domain intent induction without in-domain data

Abstract

We present our work on Track 2 in the Dialog System Technology Challenges 11 (DSTC11). DSTC11-Track2 aims to provide a benchmark for zero-shot, cross-domain, intent-set induction. In the absence of in-domain training dataset, robust utterance representation that can be used across domains is necessary to induce users' intentions. To achieve this, we leveraged a multi-domain dialogue dataset to fine-tune the language model and proposed extracting Verb-Object pairs to remove the artifacts of unnecessary information. Furthermore, we devised the method that generates each cluster's name for the explainability of clustered results. Our approach achieved 3rd place in the precision score and showed superior accuracy and normalized mutual information (NMI) score than the baseline model on various domain datasets.