Smart Virtual Agent accelerated with Transfer Learning for Customer Service (ART/342CP)

Smart Virtual Agent accelerated with Transfer Learning for Customer Service (ART/342CP)

Smart Virtual Agent accelerated with Transfer Learning for Customer Service (ART/342CP)
ART/342CP
Platform
28 / 03 / 2022 - 27 / 09 / 2023
7,851.625

Ms Adele WANG

AAHAA LIMITED (Contract Service)
SINO ESTATES MANAGEMENT LIMITED (Contract Service)


Artificial intelligence (AI), through voice dialogue to communicate with people, can turn some manual services into robotic services, such as telephone inquiry and business processing, the current AI services have some shortcomings and capabilities to be improved and enhanced. When artificial intelligence answers calls, the multilingual scenario is a tough situation that AI platform is usually facing in the service. Language recognition and semantic understanding of multi-language is an important part of the research of this project. Moreover, in this project, we will build the vertical domain knowledge for different industries, the application specific knowledge will be added into the domain knowledge so that the AI agent can understand the context of conversation during the call. How to generate an answer is another important technique after understanding the input language and its semantics. In addition, we will study the algorithm of voice cloning over text-to-speech engine and generation of virtual character, such that virtual AI agent can give human-like feedback with facial expression with specific voiceprint. These studies will enable us to build a full stack intelligent AI platform for smart AI customer services.