Speaker-change Aware CRF for Dialogue Act Classification
arXiv:2004.02913
Abstract
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.
typo fix: argmin -> argmax
References in corpus (6)
- Optimal Hyperparameters for Deep LSTM-Networks for Sequence Labeling Tasks
- TENER: Adapting Transformer Encoder for Named Entity Recognition
- Hierarchically-Refined Label Attention Network for Sequence Labeling
- Transfer Learning for Sequence Labeling Using Source Model and Target Data
- Energy-based Self-attentive Learning of Abstractive Communities for Spoken Language Understanding
- LinTO : Assistant vocal open-source respectueux des données personnelles pour les réunions d'entreprise