Dialog Router: Automated Dialog Transition via Multi-Task Learning

Abstract

Dialog Router is a general paradigm for human-bot symbiosis dialog systems to provide friendly customer care service. It is equipped with a multi-task learning model to automatically capture the underlying correlation between multiple related tasks, i.e. dialog classification and regression, and greatly reduce human labor work for system customization, which improves the accuracy of dialog transition. In addition, for learning the multi-task model, the training data and labels are easy to collect from human-to-human historical dialog logs, and the Dialog Router can be easily integrated into the majority of existing dialog systems by calling general APIs. We conduct experiments on real-world datasets for dialog classification and regression. The results show that our model achieves improvements on both tasks, which benefits the dialog transition application. The demo illustrates our method’s effectiveness in a real customer care service.

Research on Standard Cost Prediction of Intelligent Overhaul Based on Multiparticle Optimization

“…This method cannot sufficiently leverage the underlying knowledge within subtasks and various kinds of dialog data. Another similar work is a demo system for dialog transition (Huang et al, 2021) which tries a vanilla multi-task learning method (i.e., a dialog encoder tailed with two prediction subtasks). However, it still hardly fully utilizes the knowledge among dialog data and subtasks.…”

“…• Joint-CNN-BiRNN-Att (Joint-CBA) is a multi-task model by extending the last layer of CBA model to simultaneously predict NPS and dialog categories. • Vanilla Multi-task Model (VMM) follows the vanilla multi-task learning paradigm and uses the standard BERT encoder to encode only the utterance data (Huang et al, 2021), without any gated-mechanism modules.…”