Developing a Graphical User Interface for an Artificial Intelligence-Based Voice Assistant

Introduction

An Intelligent Personal Assistant (IPA) is a computer program with Artificial Intelligence (Chowdury, 2018) (AI) created to help the users in their task using Natural language processing. It has access to the large amount of information on a device or on an online platform, the helps voice assistants to perform the tasks easily. They communicate easily with users even while answering their queries and perform actions to satisfy the requests. Recently developed Voice Assistants performs wide variety of tasks from setting alarms to more complex things like taking notes or making phone calls. (Chowdury, 2018). There are three stages in which the voice assistant work: text to speech, text to intention, and intention to action. The basic six components of any voice assistant is Speech Recognition Engine, language apperception, Dialog Manager, Natural Language generator and knowledge based.

The Voice assistants provide wide variety of services such as giving information and facts from Google, Play music from streaming services such as Spotify, Wynk Music, videos and TV shows or movies on television. User can ask questions to control home automation etc. (Veena, 2018). According to Ni (2020) Intent Detection and Slot Filling joint task, as one of the core sub-tasks in NLU, has been widely used in different human–computer interaction scenarios (Ni, 2020). The Automatic Speech Recognition used by the authors is Pyttsx3, it is the text to speech conversion library in python. It also includes the TTS engines such as Sapi5, NSSS and espeak. Windows supports only Sapi5 as its driver, NSSS for Mac OS and espeak support other operating systems. Voice Assistant need not only communicate through voice but the authors have taken a step ahead by introducing Graphical User Interface (GUI). (Figure 1) The authors have integrated the GUI application using Tkinter which is a Python Toolkit, where user can interact with the AI-based voice assistant using visual components.

The initial voice assistants could do a very few tasks, but with the better NLP algorithms and more data available they can even fix appointments with the doctor or shop nearby. They can also send messages to the people as both SMS using other applications; also, more apps will be supporting integration with voice assistants in the OS without opening apps – it can make a tweet, read a recent message or send it to someone. Helping the people with disabilities either guiding them or speaking on their behalf, it becomes easier for them. The main aim of this proposal is to develop a functional IPA which performs mental tasks such as turning ON/OFF smart phone applications with the help of voice user interface (VUI) which is used to listen and process audio commands and integrating it with the Graphical User Interface.

Figure 1.

Artificial Intelligence-based Voice Assistant Integrated with Graphical User Interface

Literature Survey

Vashistha et al., (2019) used a technology called Zero UI that uses movement, statement and also introspections which make a system react to users according to the situations. Instead of depending on typing, clicking and knocking, the users have to enter the data through voice. Interactions will be changed from landlines and personal computers to the electronic gadgets which users can interact with. All these can happen with the robotics and IoT fields.

Burbach et al., (2019) in this study the author explains about the approval of voice assistant, Performance of Natural Language Processing (NLP) and Price etc., Privacy is the important factor for the acknowledgement of voice assistants. Many techniques were used to develop a voice assistant for their explored factors for the likely users, but according to this study, the author developed voice assistant for particular user groups. The acceptance of voice assistant and decision to utilize the voice assistant, relies upon the various elements of privacy which are most important other than four different groups of virtual assistants with different inclinations could be identified.