Context Does Matter: Implications for Crowdsourced Evaluation Labels in Task-Oriented Dialogue Systems.
Clemencia SiroMohammad AliannejadiMaarten de RijkePublished in: CoRR (2024)
Keyphrases
- dialogue system
- human computer
- dialogue management
- user model
- natural language
- tutorial dialogue
- mixed initiative
- natural language generation
- gold standard
- human users
- dialogue manager
- spoken dialogue systems
- language understanding
- training data
- low quality
- knowledge acquisition
- machine learning
- human computer interaction
- context aware
- cooperative