Bias Finder

Feeling unfairly graded? Bias Finder ensures fair, bias-free essay grading using AI. It detects anomalies and aligns scores with rubrics. Trust Bias Finder for accurate results!

Comment

About the Project: Bias Finder

Inspiration

As students, we’ve all experienced the frustration of receiving a grade that didn’t seem to reflect the effort or quality of our work. Whether it’s due to human error, bias, or inconsistencies in grading, the impact on a student’s morale can be significant. This inspired us to create Bias Finder – an AI-powered solution to ensure fair, unbiased, and accurate essay grading. Our goal was to empower students with transparency and confidence in their academic assessments.

What We Learned

Building Bias Finder taught us a lot about both the technical and human aspects of grading:

  • Natural Language Processing (NLP): We dove deep into text analysis to align essay content with rubric criteria.
  • The Complexity of Bias: Understanding how grading bias manifests helped us design algorithms to detect and eliminate it.
  • User Needs: We realized the importance of making the interface intuitive for both students and educators.
  • Team Collaboration: Effective teamwork and task delegation were key in delivering this project within tight hackathon timelines.

How We Built It

Tools and Technologies:

  • Frontend: We used JavaScript and React to design a clean and user-friendly interface.
  • Machine Learning: We used the OpenAI api in the back end to help analyze the inputs.
  • Deployment: - The project was hosted on Google Cloud Platform (GCP), leveraging its scalable and secure infrastructure.

Key Features:

  1. Bias-Free Grading: AI-driven analysis that ensures fair and transparent scores.
  2. Rubric Alignment: Automatically matches essay content with rubric criteria.

Challenges We Faced

  • Defining Fairness: Translating the subjective nature of grading into objective algorithms was a significant challenge.
  • Rubric Variations: Accounting for the diversity of rubrics across disciplines required flexibility in our models.
  • Time Constraints: Building a functional prototype within the hackathon’s timeframe pushed us to prioritize features and work efficiently.

Conclusion

Bias Finder is more than just a project; it’s a step toward ensuring fairness and transparency in education. We’re proud of what we’ve accomplished and excited about the potential to scale and improve the tool in the future. With Bias Finder, students can focus on learning and growth, knowing their efforts will be evaluated fairly.

Share this project:

Updates