Name
Abstract | ||
|---|---|---|
Prosody is an important cue for identifying dialog acts. In this paper, we show that modeling the sequence of acoustic- prosodic values as n-gram features with a maximum entropy model for dialog act (DA) tagging can perform better than conventional approaches that use coarse representation of the prosodic contour through acoustic correlates of prosody. We also propose a discriminative framework that exploits preceding context in the form of lexical and prosodic cues from previous discourse segments. Such a scheme facilitates online DA tagging and offers robustness in the decoding process, unlike greedy decoding schemes that can potentially propagate errors. Using only lexical and prosodic cues from 3 previous utterances, we achieve a DA tagging accuracy of 72% compared to the best case scenario with accurate knowledge of previous DA tag, which results in 74% accuracy. |
Year | DOI | Venue |
|---|---|---|
2008 | 10.1109/ICASSP.2008.4518789 | Las Vegas, NV |
Keywords | Field | DocType |
decoding,maximum entropy methods,speech processing,speech recognition,acoustic prosodic value sequence,coarse representation,decoding process,discourse segments,discriminative framework,intonation modeling,lexical cues,maximum entropy model,n-gram features,online dialog act tagging,prosodic contour,prosodic cues,dialog act tagging,discourse context,discriminative modeling,maximum entropy model,prosody | Dialog box,Dialog act,Prosody,Speech processing,Computer science,Robustness (computer science),Speech recognition,Artificial intelligence,Natural language processing,Decoding methods,Principle of maximum entropy,Discriminative model | Conference |
Volume | ISSN | ISBN |
4518789 | 1520-6149 E-ISBN : 978-1-4244-1484-0 | 978-1-4244-1484-0 |
Citations | PageRank | References |
5 | 0.49 | 1 |
Authors | ||
3 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Vivek Kumar Rangarajan Sridhar | 1 | 144 | 11.21 |
| Narayanan Shrikanth | 2 | 5558 | 439.23 |
| Srinivas Bangalore | 3 | 1319 | 157.37 |