Spaces:
Sleeping
Sleeping
Commit
·
ca80351
1
Parent(s):
7040d1c
updated readme
Browse files
README.md
CHANGED
|
@@ -1,11 +1,45 @@
|
|
| 1 |
---
|
| 2 |
-
title:
|
| 3 |
-
emoji:
|
| 4 |
colorFrom: blue
|
| 5 |
colorTo: pink
|
| 6 |
sdk: docker
|
| 7 |
-
pinned:
|
| 8 |
-
short_description:
|
| 9 |
---
|
| 10 |
|
| 11 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
title: Accessibility Bug Prediction with ALBERT
|
| 3 |
+
emoji: 🐞
|
| 4 |
colorFrom: blue
|
| 5 |
colorTo: pink
|
| 6 |
sdk: docker
|
| 7 |
+
pinned: true
|
| 8 |
+
short_description: An AI-powered model to classify bug reports as accessibility-related or not, with Jira integration.
|
| 9 |
---
|
| 10 |
|
| 11 |
+
# Accessibility Bug Prediction Using ALBERT 🐞
|
| 12 |
+
|
| 13 |
+
This project leverages the **ALBERT (A Lite BERT)** model to classify software bug reports into two categories:
|
| 14 |
+
1. Accessibility-related bugs.
|
| 15 |
+
2. Non-accessibility bugs.
|
| 16 |
+
|
| 17 |
+
It also includes a **custom Jira plugin** to integrate the AI model into the bug-tracking workflow, making it easier for development teams to identify and prioritize accessibility issues.
|
| 18 |
+
|
| 19 |
+
## Key Features ✨
|
| 20 |
+
- **State-of-the-Art NLP**: Utilizes the ALBERT transformer model, fine-tuned for high accuracy on bug report classification tasks.
|
| 21 |
+
- **Custom Dataset**: The model was trained from scratch on a dataset collected by the research team.
|
| 22 |
+
- **Jira Plugin Integration**: Seamlessly integrates the classification system into Jira to enhance accessibility compliance workflows.
|
| 23 |
+
- **Research Collaboration**: Developed under the guidance of **Professor Wajdi Aljedaani**, a UX and Human-Centered AI researcher.
|
| 24 |
+
|
| 25 |
+
## How It Works 🚀
|
| 26 |
+
1. **Input**: Provide a textual description of a bug report.
|
| 27 |
+
2. **Prediction**: The ALBERT model analyzes the text and classifies the bug as either accessibility-related or not.
|
| 28 |
+
3. **Output**: Use the results directly or integrate them into Jira for workflow optimization.
|
| 29 |
+
|
| 30 |
+
## Applications 🛠️
|
| 31 |
+
- **Software Development**: Identify accessibility bugs to ensure compliance with standards like WCAG.
|
| 32 |
+
- **Quality Assurance**: Optimize testing and prioritization for accessibility-related issues.
|
| 33 |
+
- **Research in UX and AI**: Leverage insights for designing inclusive and accessible systems.
|
| 34 |
+
|
| 35 |
+
## Deployment 🌐
|
| 36 |
+
The model is hosted on **Hugging Face Spaces**, providing an interactive and user-friendly web interface.
|
| 37 |
+
|
| 38 |
+
[Try the Model on Hugging Face](https://huggingface.co/spaces/shivamjadhav/albert_latest_96)
|
| 39 |
+
|
| 40 |
+
## About the Research 🤝
|
| 41 |
+
This project was developed as part of a research initiative at **UNT** under **Professor Wajdi Aljedaani**'s guidance. It emphasizes the intersection of AI, UX, and accessibility to drive impactful solutions for software development.
|
| 42 |
+
|
| 43 |
+
---
|
| 44 |
+
|
| 45 |
+
Check out the configuration reference at [Hugging Face Docs](https://huggingface.co/docs/hub/spaces-config-reference).
|