Batch recurrent Q-learning trained on recorded dyadic dialogues estimates higher engagement for a laugh-backchannel policy than supervised imitation, but only through off-policy evaluation without human trials.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Batch Recurrent Q-Learning for Backchannel Generation Towards Engaging Agents
Batch recurrent Q-learning trained on recorded dyadic dialogues estimates higher engagement for a laugh-backchannel policy than supervised imitation, but only through off-policy evaluation without human trials.