Robust Dialogue State Tracking with Weak Supervision and Sparse Data
This addresses the problem of sample sparsity and distributional shift in dialogue state tracking for AI systems, offering a more robust and ontology-independent approach.
The paper tackles the challenge of generalizing dialogue state tracking to new data by proposing a training strategy that eliminates the need for fine-grained manual span labels, using novel dropout methods and a unified encoder architecture to improve robustness, achieving state-of-the-art performance on benchmarks.
Generalising dialogue state tracking (DST) to new data is especially challenging due to the strong reliance on abundant and fine-grained supervision during training. Sample sparsity, distributional shift and the occurrence of new concepts and topics frequently lead to severe performance degradation during inference. In this paper we propose a training strategy to build extractive DST models without the need for fine-grained manual span labels. Two novel input-level dropout methods mitigate the negative impact of sample sparsity. We propose a new model architecture with a unified encoder that supports value as well as slot independence by leveraging the attention mechanism. We combine the strengths of triple copy strategy DST and value matching to benefit from complementary predictions without violating the principle of ontology independence. Our experiments demonstrate that an extractive DST model can be trained without manual span labels. Our architecture and training strategies improve robustness towards sample sparsity, new concepts and topics, leading to state-of-the-art performance on a range of benchmarks. We further highlight our model's ability to effectively learn from non-dialogue data.