SMART INDIA HACKATHON 2024

TITLE PAGE

Problem Statement ID – SIH1777

Statement Title- AI-Driven Inclusive

Assessment Tools for Skill Ecosystem

Theme- Smart Education

PS Category- Software

Team ID- 21603

Team Name- Byte Brigade

 Byte
Brigade
IDEA TITLE
❖ Proposed Solution (Describe your Idea/Solution/Prototype)
We have developed an AI-powered Quality Assessment named Vidya Sathi Tool to evaluate candidates across
the entire skill ecosystem, including pen-paper exams, online MCQs, descriptive exams, practical exams, and viva
voce. The prototype will function as follows:
Our approach involves creating a web-based solution for providing comprehensive assessments across
multiple formats, catering to diverse learning levels and modes (online, offline, and blended).
The platform ensures inclusivity for Persons with Disabilities (PWD) by offering voice-to-text, text-to-speech,
alternative input methods, and customizable formats to meet their needs.
Enhance learning with adaptive AI-powered personalization, which adjusts the difficulty of questions based on
candidate performance, providing tailored feedback for continuous improvement.
Empower educators and candidates with real-time analytics and detailed feedback, visualized through intuitive
dashboards for better decision-making and progress tracking.
Maintain fairness and consistency with standardized assessments across institutions, ensuring uniform
evaluation of skills and competencies.
Our multi-mode assessment system allows for flexible testing solutions, enabling candidates to participate in
exams from remote or underserved areas with online, offline, and blended options.

@SIH Idea submission- Template 2

 Byte
Brigade
TECHNICAL APPROACH
Use Cases
process flow diagram A (PWD)candidate with a disability needs to complete an assessment
using assistive technologies like voice-to-text or text-to-speech.
How the Tool Helps:
The candidate utilizes the tool’s accessibility features, such as voice-to-text
input and text-to-speech for reading questions.
The system customizes the assessment format to accommodate the
candidate’s needs, ensuring they can complete the test effectively.
This inclusivity ensures that the assessment is fair and accessible to all
candidates, regardless of their physical abilities.
An Educator Monitoring Real-Time Candidate Performance During an
Exam An educator needs to monitor multiple candidates’ performance in
real-time during a live assessment to identify trends and areas needing
intervention.
Tools & Technologies to be used : How our solution Helps:
Python, TensorFlow, JavaScript, Tailwind CSS, The educator accesses the tool's real-time analytics dashboard.
Bootstrap, React.js, Next.js ,Google Cloud The dashboard displays performance metrics, including answer accuracy
and time spent on each question.
Speech-to-Text API,AWS, PostgreSQL, Redis
Based on the live data, the educator can quickly identify candidates who
may need additional support or intervention.

@SIH Idea submission- Template 3

 Byte
Brigade
FEASIBILITY AND VIABILITY
Why is our solution both feasible and viable? What are the potential challenges faced and risks to our idea, and
what are the solutions we have implemented to counter the risks
Technical Feasibility: faced?
The technologies required for developing the AI-powered Quality Assessment Tool
include React.js for the frontend, Node.js for the backend, TensorFlow for AI/ML User Adoption and Training:
functionalities, Redis for real-time data processing, PostgreSQL/MySQL for structured Challenge: Ensuring that users (educators, candidates) adopt and effectively use
data management, and MongoDB for unstructured data. These technologies are well- the tool may require training and support.
established, widely supported, and have extensive documentation.
Solution: Develop comprehensive user guides and training materials. Offer
Technical expertise for implementing these technologies is readily available, facilitating
support through helpdesks or online forums to assist users with any issues they
the development process.
encounter.
Operational Feasibility:
Integration with Existing Systems: The tool can be integrated with existing Learning AI Model Accuracy and Bias:
Management Systems (LMS) and assessment platforms with minimal disruption, using Challenge: AI models may exhibit biases or inaccuracies that could affect the
standard APIs and data exchange formats. fairness of assessments.
A user-friendly interface designed with frameworks like React.js ensures that Solution: Use diverse training datasets and apply fairness algorithms to mitigate
educators, candidates, and administrators can use the tool effectively with basic bias. Regularly evaluate and update models to improve accuracy and address
training. potential biases.

Economic Feasibility: Accessibility Standards:

Maintenance Costs: Maintenance costs are expected to be low, involving ongoing Challenge: Meeting diverse accessibility requirements for all users, including
updates, bug fixes, and infrastructure management. Cloud services provide cost- those with disabilities, can be challenging.
efficient scaling and maintenance options. Solution: Adhere to Web Content Accessibility Guidelines and conduct regular
Potential revenue streams include subscription models for educational institutions, accessibility audits. Incorporate feedback from users with disabilities to improve
licensing fees, or government grants for improving educational assessment practices. accessibility features.

@SIH Idea submission- Template 4

 Byte
Brigade
IMPACT AND BENEFITS
Potential Impact of Vidhya Sathi on the Target Audience
USP
Enhanced Learning Outcomes:
Impact: Delivers personalized, adaptive assessments that cater to
individual skill levels, leading to improved learning outcomes.
Candidates benefit from tailored feedback that addresses their
specific strengths and weaknesses, enhancing their overall learning
experience.
Increased Accessibility and Inclusivity:
Impact: Ensures equitable access to quality assessments for
Persons with Disabilities (PWD) by incorporating accessibility
features like voice-to-text and alternative input methods. This Benefits of the solution (social, economic)
inclusivity allows all candidates, regardless of physical or cognitive Enhanced Accessibility:
limitations, to participate fully in the assessment process. Ensures that candidates, including Persons with Disabilities (PWD), have equal
Informed Decision-Making: opportunities through accessible assessments tailored to various needs (e.g., voice-to-
Impact: Provides educators and trainers with real-time analytics text, text-to-speech).
and detailed feedback, enabling them to make data-driven Adaptive Assessments:
decisions to refine educational programs and training modules. This Provides personalized assessments that adjust based on individual performance,
leads to more effective teaching strategies and improved helping to identify specific strengths and areas for improvement.
educational outcomes. Comprehensive Coverage:
Standardization and Quality Assurance: Offers a wide range of assessment types, supporting a holistic view of a candidate’s
Impact: Ensures consistency in assessment practices across abilities.
different regions and institutions by standardizing evaluation Standardized Assessments:
criteria. This contributes to a reliable and credible skill ecosystem, Ensures consistency in evaluations across different regions and institutions, promoting
maintaining high standards of quality and rigor in skill assessments. fairness in assessing skills and competencies

@SIH Idea submission- Template 5

 Byte
Brigade
RESEARCH AND REFERENCES
1. https://www.khanacademy.org
2. https://www.coursera.org
3. Brusilovsky, P., & Millán, E. (2007). User Models for Adaptive Hypermedia and Adaptive Educational
Systems. The Adaptive Web. Springer.
4. Wang, M., & Heffernan, N. (2018). Improving Student Learning Through Adaptive Learning Systems.
Computers & Education, 123, 254-263.
link(https://www.sciencedirect.com/science/article/abs/pii/S0360131517302789)
5. Makar, K., & Lee, M. (2017). Real-Time Learning Analytics and Feedback Systems. Learning Analytics Review,
5(1), 10-25(https://www.researchgate.net/publication/319133069_Review_of_Research_on_Student-
Facing_Learning_Analytics_Dashboards_and_Educational_Recommender_Systems)

@SIH Idea submission- Template 6