CS224s class notes End-to-end speech

CS224s class notes: End-to-end speech ============= IN CS224S ==================== Emotional signals that are difficult to fake:      smiles, blushing, 

—> Speed-dating: flirting women use more hedges (—> creak)

Drahota, A., Costall, A., & Reddy, V. (2008). The vocal communication of different kinds of smile. Speech Communication, 50(4), 278-287.

Honest signals (Pentland,  2008)   Brunswikian Lens

Liscombe et al 2003 — better features.  Hirschberg, J., Liscombe, J., & Venditti, J. (2003). Experiments in emotional speech. In ISCA & IEEE Workshop on Spontaneous Speech Processing and Recognition.

    MFCC is not a good feature extractor for emotional speech. (see slides)

In a lot of emotion detection work, they are working with small datasets, hand-label centers of vowels. 

Ang, J., Dhillon, R., Krupski, A., Shriberg, E., & Stolcke, A. (2002, September). Prosody-based automatic detection of annoyance and frustration in human-computer dialog. In INTERSPEECH.

Lecture 13, slides 46-48: Good example of analysis he hopes for. 

Goodwin, Jurafsky