Joint System-Wise Optimization for Pipeline Goal-Oriented Dialog System
This work addresses the challenge of optimizing entire dialog systems rather than individual components, which is crucial for developers of conversational AI, though it is incremental as it builds on prior system-wise evaluation methods.
The paper tackled the problem of improving system-wise performance in pipeline goal-oriented dialog systems by proposing joint optimization techniques, resulting in a 12% increase in success rate in automatic evaluation and a 16% increase in human evaluation on the MultiWOZ 2.1 benchmark.
Recent work (Takanobu et al., 2020) proposed the system-wise evaluation on dialog systems and found that improvement on individual components (e.g., NLU, policy) in prior work may not necessarily bring benefit to pipeline systems in system-wise evaluation. To improve the system-wise performance, in this paper, we propose new joint system-wise optimization techniques for the pipeline dialog system. First, we propose a new data augmentation approach which automates the labeling process for NLU training. Second, we propose a novel stochastic policy parameterization with Poisson distribution that enables better exploration and offers a principled way to compute policy gradient. Third, we propose a reward bonus to help policy explore successful dialogs. Our approaches outperform the competitive pipeline systems from Takanobu et al. (2020) by big margins of 12% success rate in automatic system-wise evaluation and of 16% success rate in human evaluation on the standard multi-domain benchmark dataset MultiWOZ 2.1, and also outperform the recent state-of-the-art end-to-end trained model from DSTC9.