UniDU: Towards A Unified Generative Dialogue Understanding Framework
This work addresses the lack of shared knowledge across dialogue understanding tasks, offering a unified approach that improves performance and generalization, though it is incremental in building on existing pre-trained models.
The paper tackles the problem of isolated models for different dialogue understanding tasks by proposing UniDU, a unified generative framework that reformulates tasks into a prompt-based paradigm and introduces a multi-task training strategy, achieving superior performance on ten datasets across five tasks.
With the development of pre-trained language models, remarkable success has been witnessed in dialogue understanding (DU). However, current DU approaches usually employ independent models for each distinct DU task without considering shared knowledge across different DU tasks. In this paper, we propose a unified generative dialogue understanding framework, named {\em UniDU}, to achieve effective information exchange across diverse DU tasks. Here, we reformulate all DU tasks into a unified prompt-based generative model paradigm. More importantly, a novel model-agnostic multi-task training strategy (MATS) is introduced to dynamically adapt the weights of diverse tasks for best knowledge sharing during training, based on the nature and available data of each task. Experiments on ten DU datasets covering five fundamental DU tasks show that the proposed UniDU framework largely outperforms task-specific well-designed methods on all tasks. MATS also reveals the knowledge-sharing structure of these tasks. Finally, UniDU obtains promising performance in the unseen dialogue domain, showing the great potential for generalization.