MyAnswer

MyAnswer is an AI web application designed to enhance the exam experience for partially blind individuals. Platform for remote assessments receive instant feedback through audio and text.

Comment

MyAnswer 🎤

MyAnswer is an innovative web application designed to enhance the viva voce experience for partially blind individuals. Leveraging the power of Vertex AI and generative AI, this app provides a seamless and effective platform for remote examinations and assessments. Teachers can create and manage tests, while students can participate in exams and receive instant feedback through audio and text.

Team Members

  1. Sandeep Sreekumar
  2. Nevin
  3. Jebin Shaju
  4. Aleena Maria Rajesh

Link to Product Walkthrough

Link to video

The site is live at MyAnswer.

Inspiration

The inspiration for MyAnswer came from the desire to make the education process more inclusive. We recognized the challenges faced by partially blind individuals during oral examinations and sought to create a solution that could bridge this gap using modern AI technologies. Our goal was to develop a platform that not only aids in conducting viva voce but also enhances the learning experience by providing immediate, constructive feedback.

What It Does

MyAnswer facilitates remote viva voce examinations for partially blind students, enabling them to answer questions orally. It uses speech-to-text technology to transcribe responses and generative AI to evaluate these answers. Teachers can create and manage tests, view student submissions, and provide feedback. The platform offers an intuitive interface for both teachers and students, ensuring a smooth and accessible examination process.

How We Built It

We built MyAnswer using a combination of cutting-edge technologies and frameworks:

  • Streamlit for building the web interface, offering a simple and interactive user experience.
  • Pyrebase for integrating Firebase, which handles authentication, database management, and storage.
  • Vertex AI for leveraging advanced AI capabilities to evaluate student answers and generate feedback.
  • gTTS (Google Text-to-Speech) for converting text feedback into audio, aiding partially blind users.
  • TemporaryFile and Base64 for handling file operations and encoding.
  • Re (Regular Expressions) for processing and validating input data.
  • UUID and JSON for generating unique identifiers and managing data interchange.

Vertex AI and Customized Recommendations

Leveraging Vertex AI

Vertex AI plays a crucial role in MyAnswer by providing advanced AI capabilities to evaluate student responses and generate feedback. This integration allows the platform to:

  • Analyze Responses: Vertex AI analyzes student answers to understand their accuracy and depth.
  • Generate Feedback: The AI generates detailed feedback, highlighting areas of strength and improvement.
  • Provide Recommendations: Based on the analysis, Vertex AI offers customized learning recommendations to help students improve their understanding and performance.

Customized Learning Recommendations

Vertex AI not only evaluates student answers but also helps tailor educational experiences to individual needs:

  • Identify Knowledge Gaps: The AI identifies specific areas where a student may have gaps in understanding.
  • Suggest Resources: It recommends relevant study materials, videos, and articles to help students learn more about the topics they find challenging.
  • Track Progress: The platform tracks student progress over time, adjusting recommendations to ensure continuous improvement.
  • Personalized Learning Paths: By analyzing performance data, Vertex AI creates personalized learning paths that guide students through their educational journey.

Challenges We Ran Into

Building MyAnswer was not without its challenges:

  • Speech Recognition Accuracy: Ensuring the speech-to-text functionality accurately captured diverse accents and speech patterns was a significant hurdle.
  • AI Feedback Precision: Developing a generative AI model that could provide precise and helpful feedback required extensive training and fine-tuning.
  • Accessibility: Designing an interface that was truly accessible for partially blind users involved rigorous testing and iterations.
  • Integration: Seamlessly integrating various technologies (Firebase, Vertex AI, Streamlit) required meticulous planning and execution.

Accomplishments That We're Proud Of

  • Successfully creating a platform that enhances the examination experience for partially blind students.
  • Integrating advanced AI technologies to provide immediate, detailed feedback to students.
  • Developing an intuitive and accessible user interface that caters to both teachers and students.
  • Receiving positive feedback from initial user testing, validating the effectiveness and usability of the platform.

What We Learned

Throughout the development of MyAnswer, we gained valuable insights into:

  • The importance of accessibility in educational tools and how technology can bridge gaps.
  • The intricacies of speech recognition and the challenges involved in developing accurate AI models.
  • The power of generative AI in providing personalized feedback and enhancing the learning experience.
  • Effective ways to integrate multiple technologies to create a cohesive and functional application.

What's Next for MyAnswer

Looking ahead, we plan to:

  • Enhance AI Capabilities: Further improve the generative AI model to provide more detailed and personalized feedback.
  • Expand Accessibility Features: Incorporate additional accessibility features, such as support for more languages and enhanced audio feedback options.
  • Mobile Application: Develop a mobile version of MyAnswer to increase accessibility and convenience for users.
  • Collaborations and Partnerships: Partner with educational institutions and organizations to broaden the reach and impact of MyAnswer.
  • Continuous User Feedback: Gather ongoing feedback from users to continuously improve the platform and address any emerging needs.

Key Features

Teacher Functionality:

  • Sign Up and Log In: Teachers can create an account and log in to access the dashboard.
  • Create Tests: Teachers can create tests by adding questions and answers.
  • Manage Questions: Teachers can view and manage the questions for each test.
  • View Submissions: Teachers can view student submissions and check their marks.
  • Generative AI Evaluation: The app uses generative AI to evaluate student answers and provide detailed feedback.

Student Functionality:

  • Sign Up and Log In: Students can create an account and log in to access the dashboard.
  • Enter Examination Code: Students can enter examination codes to start their exams.
  • Speech-to-Text Answer Capture: The app uses speech-to-text to capture student answers.
  • Audio Feedback: Students receive audio feedback on their answers.
  • View Results: Students can view their previous exam results and receive instant feedback.

Libraries Used

  • Streamlit
  • Pyrebase
  • Vertex AI
  • gTTS (Google Text-to-Speech)
  • TemporaryFile
  • Base64
  • Re
  • UUID
  • JSON

Configuration Steps

  1. Clone the Repository:

    git clone https://github.com/sandeepsreekumar4067/StreamLit_Tinkerhub_2024
cd myanswer

  2. Install Required Libraries:

    pip install -r requirements.txt

  3. Firebase Configuration:

    • Set up your Firebase project and configure the firebaseConfig in the app code. python firebaseConfig = { 'apiKey': "YOUR_API_KEY", 'authDomain': "YOUR_AUTH_DOMAIN", 'projectId': "YOUR_PROJECT_ID", 'storageBucket': "YOUR_STORAGE_BUCKET", 'messagingSenderId': "YOUR_MESSAGING_SENDER_ID", 'appId': "YOUR_APP_ID", 'measurementId': "YOUR_MEASUREMENT_ID", 'databaseURL': 'YOUR_DATABASE_URL' }

  4. Vertex AI Initialization:

    • Initialize Vertex AI with your project details. python vertexai.init(project="YOUR_PROJECT_ID", location="YOUR_LOCATION")

Running the Application

  1. Run the Streamlit App:

    streamlit run app.py

  2. Access the App:

Deployment on Google Cloud

  1. Build the Docker Image:

    docker build -t myanswer .

  2. Push the Docker Image to Google Container Registry:

    docker tag myanswer gcr.io/your-project-id/myanswer
docker push gcr.io/your-project-id/myanswer

  3. Deploy to Cloud Run:

    gcloud run deploy myanswer --image gcr.io/your-project-id/myanswer --platform managed

Conclusion

MyAnswer revolutionizes the viva voce process for partially blind individuals by integrating advanced AI technologies to create an inclusive, effective, and user-friendly examination platform. By leveraging the power of Vertex AI, MyAnswer not only provides immediate feedback but also offers personalized learning recommendations, making the educational process more adaptive and impactful. Join us in making education more accessible and engaging for everyone.

Share this project:

Updates