ASSIST: Towards Label Noise-Robust Dialogue State Tracking
This addresses the challenge of noisy annotations in dialogue state tracking for researchers and practitioners, offering a practical solution without costly re-annotation, though it is incremental as it builds on existing methods for noise robustness.
The paper tackles the problem of label noise in dialogue state tracking datasets like MultiWOZ, proposing the ASSIST framework to train models robustly from noisy labels, resulting in improvements of up to 28.16% in joint goal accuracy on MultiWOZ 2.0 and 8.41% on MultiWOZ 2.4 compared to using only noisy labels.
The MultiWOZ 2.0 dataset has greatly boosted the research on dialogue state tracking (DST). However, substantial noise has been discovered in its state annotations. Such noise brings about huge challenges for training DST models robustly. Although several refined versions, including MultiWOZ 2.1-2.4, have been published recently, there are still lots of noisy labels, especially in the training set. Besides, it is costly to rectify all the problematic annotations. In this paper, instead of improving the annotation quality further, we propose a general framework, named ASSIST (lAbel noiSe-robuSt dIalogue State Tracking), to train DST models robustly from noisy labels. ASSIST first generates pseudo labels for each sample in the training set by using an auxiliary model trained on a small clean dataset, then puts the generated pseudo labels and vanilla noisy labels together to train the primary model. We show the validity of ASSIST theoretically. Experimental results also demonstrate that ASSIST improves the joint goal accuracy of DST by up to $28.16\%$ on MultiWOZ 2.0 and $8.41\%$ on MultiWOZ 2.4, compared to using only the vanilla noisy labels.