Xiyuan Gao, MA

My research focuses on tackling the complexities of non-literal language (e.g., implications, figurative language), with a particular emphasis on sarcasm, within the field of language technology through machine learning approaches. I investigate the intersection of linguistics and language technology, employing a multimodal approach (text, audio, visual) to enhance and advance state-of-the-art methods. In addition to sarcasm recognition, my work addresses challenges related to data scarcity, where I explore innovative strategies in data augmentation and transfer learning to boost performance in resource-constrained environments. Furthermore, I am also interested in sarcastic speech synthesis, extending my research to the generation of non-literal language in both textual and spoken forms.