Enabling a Massive Data Collection for Hotel Receptionist Chatbot Using a Crowdsourcing Information System

With the heavy utilization of deep learning, modern chatbot systems require massive training dataset to achieve their peak performance. Without exception, the development of chatbot for the hotel receptionist tasks also demands a massive training dataset. Unfortunately, laborious effort and hefty time are needed to manually collect a dataset with that size. Therefore, an information system is needed to assist the dataset collection. In regards to solve the problem, this research contributes in developing the information system that enables a massive data collection for training a hotel receptionist chatbot in a reasonable time. In detail, the contribution of this research are twofold: (1) developing a realiable crowdsourcing information system design to collect and store the dataset for training hotel receptionist chatbot; and (2) developing a user experience that can stimulate the crowdworkers to provide more data.