AIDRIN: AI Data Readiness Inspector
AI Data Readiness Inspector (AIDRIN) is a framework designed to assess the readiness of datasets for AI model training. It provides a structured evaluation across key dimensions that influence the success and reliability of AI systems.
AIDRIN includes a standalone user interface that allows users to upload their data and receive detailed evaluations across six core pillars of AI data readiness:
- Data Quality
- Data Understandability and Usability
- Data Governance
- Impact of Data on AI Outcomes
- Fairness and Bias
- Privacy and Compliance
Each pillar is assessed using established metrics from academic and industry literature, helping users identify potential issues before model development begins.
In addition to the standalone AIDRIN, it is integrated with privacy-preserving federated learning workflows through a collaboration with the APPFL framework. This integration enables secure, decentralized data readiness assessments without compromising data privacy.
📄 Documentation and code for this integration are available here.
Publications​
Authors | Title | Venue | Type | Date | Links |
|---|---|---|---|---|---|
| , , | Data Readiness for AI: A 360-Degree Survey | ACM Comput. Surv. 57, 9, Article 219, 39 pages | Journal | April, 2025 | |
| , , , | AI Data Readiness Inspector (AIDRIN) for Quantitative Assessment of Data Readiness for AI | SSDBM '24: Proceedings of the 36th International Conference on Scientific and Statistical Database Management. Article No.: 7, Pages 1 - 12 | Conference | August, 2024 | |
| , , , , | CADRE: Customizable Assurance of Data Readiness in Privacy-Preserving Federated Learning | IEEE e-Science 2025 | Conference | 2025 |