DORIC : Domain Robust Fine-Tuning for Open Intent Clustering through Dependency Parsing
This work addresses the challenge of robust intent clustering across domains for dialog systems, though it appears incremental as it builds on existing fine-tuning and parsing techniques.
The authors tackled the problem of zero-shot cross-domain intent clustering without in-domain training data by fine-tuning a language model on multi-domain dialogue data and extracting Verb-Object pairs to remove unnecessary information, achieving 3rd place in precision and superior accuracy and NMI scores compared to baselines on various domain datasets.
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.