I-Hung Hsu, a computer science Ph.D. student at the USC Viterbi School of Engineering, was recently named an Amazon Alexa Graduate Fellow, a prestigious honor awarded to just 10 talented doctoral students from universities around the globe. His research on conversational artificial intelligence promises to improve and streamline the interaction between people and computers using voice.
“I’m very excited to win this award which will give me a chance to collaborate with other human language technology students,” Hsu said. “This will facilitate me to get involved with more research and topics that others are interested in.”
Amazon recently founded the award for postgraduate students specializing in conversational AI. Amazon itself is pioneering this conversational intelligence technology with the product Alexa, a voice-activated personal assistant that allows consumers to play music, check the weather and receive news, simply by speaking aloud. On its website, the company said it created the new graduate fellowship because it believes “technology innovation and entrepreneurship starts on college campuses.”
The award includes funding for research students as well as a monthly stipend to cover expenses.
Hsu received the fellowship for his work on developing a more robust speech recognition model, particularly one that can perform better in noisy environments. “Through our current achievements in speech recognition, such as Amazon Echo, Google Assistant, and Apple Siri, we know that their recognition rates are pretty decent,” Hsu said. “However, if you put your Echo in a more noisy environment, such that you are holding a party at your home, the performance would degrade.”
Human language technology, or HLT, is an emerging field with a new certification that USC Viterbi began offering this fall. HLT encompasses all new technologies that involve the integration of electronics and language, such as speech recognition, speech generation, grammar, spell-checking, understanding of different languages, and the application of auditory and typed commands.
“Unlike menus, touchscreens, or mouse clicks, using our voice to have a conversation is one of the most natural ways to use a computer,” the Amazon website noted. “It requires no learning curve.” HLT strives to make using technology even easier and more accessible to everyone.
Nanyun Peng, USC Viterbi research assistant professor of computer science and Hsu’s faculty advisor, said she “couldn’t be prouder about I-Hung winning the Alexa Graduate Fellow. It’s well-deserved.”
Hsu is excited to apply the funds received from the award to his future research, which is still in the planning stages.
“My postgraduate life has just begun,” he said. “So, my first goal is to do well on my research and keep finding research topics that are fun and engaging to me and beneficial to human society.”