LOGIN TO YOUR ACCOUNT

Username
Password
Remember Me
Or use your Academic/Social account:

CREATE AN ACCOUNT

Or use your Academic/Social account:

Congratulations!

You have just completed your registration at OpenAire.

Before you can login to the site, you will need to activate your account. An e-mail will be sent to you with the proper instructions.

Important!

Please note that this site is currently undergoing Beta testing.
Any new content you create is not guaranteed to be present to the final version of the site upon release.

Thank you for your patience,
OpenAire Dev Team.

Close This Message

CREATE AN ACCOUNT

Name:
Username:
Password:
Verify Password:
E-mail:
Verify E-mail:
*All Fields Are Required.
Please Verify You Are Human:
fbtwitterlinkedinvimeoflicker grey 14rssslideshare1
Casanueva, I.; Hain, T.; Nicolao, M.; Green, P. (2016)
Languages: English
Types: Other
Subjects:
The generalisation of dialogue state tracking\ud to unseen dialogue states can be very\ud challenging. In a slot-based dialogue system,\ud dialogue states lie in discrete space\ud where distances between states cannot be\ud computed. Therefore, the model parameters\ud to track states unseen in the training\ud data can only be estimated from more general\ud statistics, under the assumption that\ud every dialogue state will have the same underlying\ud state tracking behaviour. However,\ud this assumption is not valid. For example,\ud two values, whose associated concepts\ud have different ASR accuracy, may\ud have different state tracking performance.\ud Therefore, if the ASR performance of the\ud concepts related to each value can be estimated,\ud such estimates can be used as general\ud features. The features will help to relate\ud unseen dialogue states to states seen\ud in the training data with similar ASR performance.\ud Furthermore, if two phonetically\ud similar concepts have similar ASR\ud performance, the features extracted from\ud the phonetic structure of the concepts can\ud be used to improve generalisation. In\ud this paper, ASR and phonetic structurerelated\ud features are used to improve the\ud dialogue state tracking generalisation to\ud unseen states of an environmental control\ud system developed for dysarthric speakers.\ud

Share - Bookmark

Cite this article