DialogConv: A Lightweight Fully Convolutional Network for Multi-view Response Selection
This work addresses efficiency issues in dialogue systems for applications requiring real-time response, though it is incremental as it builds on existing convolutional approaches.
The paper tackles the problem of slow inference and large model sizes in retrieval-based dialogue systems by proposing DialogConv, a lightweight fully convolutional network for multi-view response selection, which achieves competitive effectiveness while being on average 8.5x smaller and up to 79.39x faster on CPU compared to state-of-the-art baselines.
Current end-to-end retrieval-based dialogue systems are mainly based on Recurrent Neural Networks or Transformers with attention mechanisms. Although promising results have been achieved, these models often suffer from slow inference or huge number of parameters. In this paper, we propose a novel lightweight fully convolutional architecture, called DialogConv, for response selection. DialogConv is exclusively built on top of convolution to extract matching features of context and response. Dialogues are modeled in 3D views, where DialogConv performs convolution operations on embedding view, word view and utterance view to capture richer semantic information from multiple contextual views. On the four benchmark datasets, compared with state-of-the-art baselines, DialogConv is on average about 8.5x smaller in size, and 79.39x and 10.64x faster on CPU and GPU devices, respectively. At the same time, DialogConv achieves the competitive effectiveness of response selection.