Digital Signal Processing Reference
In-Depth Information
Figure 1-6.
Sample of the corpus with dependency tags.
5.
DISCUSSION ON IN-CAR SPEECH CORPUS
In this section‚ we will be studying the characteristics of our multi-layered
in-car speech corpus of Figure 1-7. In particular‚ we will explore the
relationship between an intention in a given utterance and the utterance
length‚ and the relationship between the intentions and the associated
linguistic phenomenon. Especially‚ we will be comparing the driver's
conversations with another person (human navigator) and the human-WOZ
dialogues.
5.1
Dialogue Sequence Viewer
To understand and analyze the dialogue corpus intuitively‚ we have
constructed a dialogue sequence viewer as depicted in Figure 1-8. For this
task‚ we have also formed “turns” from speech units or tagged units‚ to
indicate the event of a speaker change. In this figure‚ each node has a tag with
a turn number‚ and the link between any two nodes implies the sequence of
the event in a given conversation. As expected‚ each turn could have more
than one LIT tag. The thickness of a link is associated with the occurrence
count of a given tag's connections. For instance‚ there are only four turns in
Figure 1-8 of the dialogue segment of Figure 1-4. We have observed that the
average turn count in the restaurant query task is approximately 10.
