论文标题
说话者更改对话ACT分类的CRF
Speaker-change Aware CRF for Dialogue Act Classification
论文作者
论文摘要
对话法案（DA）分类的最新工作将任务作为序列标记问题将其方法方法与有条件的随机场（CRF）作为最后一层的神经网络模型。 CRF在给定输入话语序列的情况下对目标DA标签序列的条件概率进行建模。但是，该任务涉及另一个重要的输入序列，即说话者的任务，这是先前工作所忽略的。为了解决这一限制，本文提出了对CRF层的简单修改，该层将扬声器变换考虑在内。 SWDA语料库上的实验表明，我们修改的CRF层的表现优于原始层，某些DA标签的边距非常宽。此外，可视化表明，我们的CRF层可以学习有意义的，复杂的过渡模式，以端到端的方式以扬声器变化为条件。代码公开可用。
Recent work in Dialogue Act (DA) classification approaches the task as a sequence labeling problem, using neural network models coupled with a Conditional Random Field (CRF) as the last layer. CRF models the conditional probability of the target DA label sequence given the input utterance sequence. However, the task involves another important input sequence, that of speakers, which is ignored by previous work. To address this limitation, this paper proposes a simple modification of the CRF layer that takes speaker-change into account. Experiments on the SwDA corpus show that our modified CRF layer outperforms the original one, with very wide margins for some DA labels. Further, visualizations demonstrate that our CRF layer can learn meaningful, sophisticated transition patterns between DA label pairs conditioned on speaker-change in an end-to-end way. Code is publicly available.
