Multipurpose Virtual Assistant Using Machine Learning

Abstract

— Virtual assistant is a software agent who will perform a different task or work for personally supported Questions or commands called a Virtual Assistant with the intelligence of a human. In some phenomena, virtual assistants can interpret human speech, and sign language responds via synthesized voices and moves to convert voice to text. Clients can interrogate with their virtual assistants to control home automation devices manage their tasks like email, to-do lists, and media playback, check weather updates interact with help of sign language with verbal command and response. Several research papers regarding multipurpose virtual assistant were reviewed and an appropriate solution to address this problem was concluded. You Only Look Once Natural Language Processing and Convolutional Neural Network (CNN) algorithms assist to perform operations with voice and text commands.  By using this approach, the final result is expected to have an accuracy score of over 96%.