Rae Koh

Rae Koh is a PhD candidate at the Interdisciplinary Graduate Programme (Global Asia), Nanyang Technological University. She is interested in how people talk and how computers can understand how people talk. Her research focuses on the prosody of emotional speech from multilingual speakers for applications in speech emotion recognition. 

Emotional expressions of bilingual speakers in spontaneous speech: Evidence from Singapore English

Following significant advancements made in the field of automatic speech recognition, the next milestone in developing computer applications that can understand human speech is speech emotion recognition (SER), that is, detecting emotional cues in speech signals. Much of the progress in the field so far has been spurred on by modelling emotions from acted data, primarily from speakers in Western societies. Little is currently understood of emotions occurring in natural conversations; even less when considering how expressions of emotions differ among multicultural and multilingual speakers in Asia. In this talk, I present findings of the various ways in which anger and happiness are communicated in spontaneous speech obtained from dyadic audio recordings of Singapore English speakers of Chinese, Malay, or Indian ethnicities engaging in a co-operative game. The implications of the findings on SER research are also discussed.