Chatbot companion

Abstract

The prevalence of loneliness among the elderly population is a significant concern, with adverse effects on mental and physical health. In Singapore, over 30% of the elderly reportedly face severe loneliness issues, a problem exacerbated by the aging population. While the government has introduced various initiatives over the years, the issue persists. However, advancements in technologies such as Large-Language Models and Speech-to-Text present new opportunities. By leveraging these technologies, we can develop a proactive and audio-driven chatbot companion to engage in meaningful conversations with the elderly. This project explores the development of such a companion, utilising technologies like Single Sign-On options such as Singpass, AI Singapore The Speech Lab's trilingual Speech-to-Text engine, Yi-34B-Chat Large-Language Model, and Microsoft Azure's human-like avatar for lip-syncing of real-time output transcription. Additionally, features such as real-time data handling for stock prices, weather updates, past lottery results, and news, automatic lottery results broadcast, along with an in-built events calendar, wake word and sleep word functionality, and the ability to initiate conversations at customisable timings, were incorporated to create a comprehensive and user-friendly application.