Voice - Based Virtual Assistant For Windows (ZIVA)

Sonam, Shubham Patel, Vaishnav Srivastav

Introduction: Voice Based Assistant Application: Voice Assistant Speech To Text Recognition for Virtual Asssitant Challenges in of Brain Cancer Detection using ML
Users can speak naturally to their Windows virtual assistant, Data Availability: Access to high-quality, voice datasets is
The implemented project is a digital and virtual life assistant, with Voice based Virtual assistant used in a wide area of issuing commands or asking questions, and the assistant will essential for training
Artificial Intelligence. It can perform the different tasks without applications like: transcribe and interpret the spoken words. the machine learning model, and data collection and curation
any physical intervention with keyboard and mouse.It can open up Smart Home Integration It is used to understand the voice of computer ad convert them to may pose challenges.
the application (if it’s installed in the system), using speech Entertainment and Media perform any required task. Model Integration: The machine learning model should be
recognition. It can process the data as per the need or can add the Office and Business Applications seamlessly integrated into the web
functionality, depends upon how we code things.
Education application, requiring careful design for efficient interaction.
Virtual assistants use natural language processing (NLP) to match
Healthcare Model Training and Updates: Procedures for training,
user text or voice input to executable commands. Many continually
learn using artificial intelligence techniques including machine Development and Programming versioning, and updating the model
learning. ZIVA, a voice-based virtual assistant, employs NLP to must be established to maintain accuracy and adapt to new data.
interpret user voice input, executing tasks and learning through AI Interpretability: In the medical field, model interpretability and
and machine learning. Methodology: Voice base Virtual Assistant explainability are crucial for
user trust and understanding
The scope of a virtual assistant for Windows using NLP is Fig:6 Speech Recognition
extensive, encompassing user interaction via natural language, Conclusion
task automation, information retrieval, system integration,
entertainment, and accessibility features. Our proposed approach, centered around a meticulously selected
NLP architecture, represents a significant stepforward in this
field. Customizing the chosen NLP model and applying
advanced data processing techniques are central to achieving
higher accuracy and robustness. These strategies, as discussed in
the literature review, are vital for overcomingchallenges related
to limited datasets and the intricate natureof voice assistant .
The evaluation of our approach, performed across diverse
datasets and using comprehensive evaluation metrics, instills
confidence in its effectiveness.
Fig.1 Voice Based Assistant Through a comparative analysis, we have observed the potential
Fig.7 System Design benefits of our approach in enhancing voice basedvirtual
Fig. 4 Flowchar assistant for windows and classification accuracy. Itsadaptability
Speech Recognition : It is used to understand the voice of
computer ad convert them to perform any required task. to different datasets and imaging modalities, aswell as its
Natural Language Processing (NLP): insights into the impact of hyperparameters andtraining
NLP enables people to ask inquiries about a subject and receive strategies, further highlight its significance.
prompt, straightforward answers. Research Gap, Objectives, Motivation & Challenges
References
NLP responds to your inquiries with precision. In other words, it Research Gap
doesn't offer extraneous or needless information. NLP facilitates Limited Data: Need larger and more diverse datasets for [1]Goh, A.S.Y., Wong, L.L. and Yap, K.Y.L., 2021. Evaluation of
human-computer communication in that language. training. COVID-19 information provided by digital voice assistants.
NLP enables these assistants to understand, interpret, and respond Handling Imbalanced Data: Address issues with unequal classes International Journal of Digital Health, 1(1).
to user queries and commands in a way that mimics human for accurate results. [2] Duque, M., Pink, S., Strengers, Y., Martin, R. and Nicholls, L.,
language comprehension 2021. Automation, wellbeing and digital voice assistants: Older
Lack of Pre-Trained Models: Training a model requires a lot of
time and resources, having pre-trained weights helps with people and Google devices. Convergence, 27(5), pp.1189-1206.
Fig.2 SR achieving better accuracy. [3] Ewers, K., Baier, D. and Höhn, N., 2020. Siri, Do I Like You?
Dataset: Objective Digital voice assistants and their acceptance by consumers.
This dataset is a simplified version of an original dataset Seamless Interaction:- Enable users to interact with their SMR-Journal of Service Management Research, 4(1), pp.52-68.
containing audio recordings from multiple speakers. Recordings Windows devices naturally using voice commands, reducing the [4] S. Subhash, P. N. Srivatsa, S. Siddesh, A. Ullas and B.
contain individual voice command for a virtual assistant. Speakers reliance on manual input methods Santhosh, "Artificial Intelligence-based Voice Assistant," 2020
vary in language, fluency level and other properties. Task Automation :-Automate routine tasks and commands to Fourth World Conference on Smart Trends in Systems, Security
Dataset of audio files containing voice commands for a generic streamline user workflows and improve overall efficiency. and Sustainability (WorldS4), London, UK, 2020,
virtual assistant. [5]Ahmad, I., Akter, T., Buher, Z., Farzan, R., Kapadia, A. and Lee,
A.J., 2022. Tangible Privacy for Smart Voice Assistants:
Motivation
Bystanders' Perceptions of Physical Device Controls. Proceedings
Voice-based interfaces cater to users with varying physical of the ACM on Human-Computer Interaction, 6(CSCW2), pp.1-31.
abilities, making technology more accessible
[6] Gamal Bohouta, Veton Z Këpuska,(2021) Next-Generation of
Voice-based virtual assistants empower elderly individuals to Virtual Personal Assistants (Microsoft Cortana, Apple Siri,
Fig. 5 NLP
engage with technology. Amazon Alexa and Google Home)IEEE CCWC 2021, The 8th
Fig.3 Dataset Voice-based interfaces cater to linguistic diversity. IEEE Annual Computing and Communication Workshop and
Voice-based virtual assistants promote a digital-first mindset. ConferenceAt: University of Nevada, Las Vegas, USA

RESEARCH POSTER PRESENTATION DESIGN © 2019

www.PosterPresentations.com