Abstract
Dialogue act recognition and simulation are traditionally considered separate processes. Here, we argue that both can be fruitfully treated as interleaved processes within the same probabilistic model, leading to a synchronous improvement of performance in both. To demonstrate this, we train multiple Bayes Nets that predict the timing and content of the next user utterance. A specific focus is on providing support for barge-ins. We describe experiments using the Let's Go data that show an improvement in classification accuracy (+5%) in Bayesian dialogue act recognition involving barge-ins using partial context compared to using full context. Our results also indicate that simulated dialogues with user barge-in are more realistic than simulations without barge-in events.
Original language | English |
---|---|
Title of host publication | 2013 IEEE Workshop on Automatic Speech Recognition and Understanding |
Place of Publication | New York |
Publisher | IEEE |
Pages | 102-107 |
Number of pages | 6 |
Publication status | Published - 2013 |
Event | 2013 IEEE Workshop on Automatic Speech Recognition and Understanding - Olomouc, United Kingdom Duration: 8 Dec 2013 → 13 Dec 2013 |
Conference
Conference | 2013 IEEE Workshop on Automatic Speech Recognition and Understanding |
---|---|
Abbreviated title | ASRU 2013 |
Country/Territory | United Kingdom |
City | Olomouc |
Period | 8/12/13 → 13/12/13 |
Keywords
- spoken dialogue systems
- dialogue act recognition
- dialogue simulation
- Bayesian nets
- barge-in
- USER SIMULATION