Skip to content

Latest commit

 

History

History
41 lines (20 loc) · 3.58 KB

File metadata and controls

41 lines (20 loc) · 3.58 KB

Module 3 Resources

Learning from Data

  • MIT CSAIL · Data-Centric AI (lecture notes/overview). Frames model-centric vs data-centric workflows and lays out a practical recipe for improving data quality through iteration, labeling standards, and targeted error analysis. https://datacentricai.org/

  • CACM: The Principles of Data-Centric AI. A quotable set of principles for dataset design, versioning, labeling consistency, augmentation, and evaluation that shift focus from models to data. https://cacm.acm.org/

  • ArXiv survey: Data-Centric Artificial Intelligence. A broad taxonomy and literature map covering data assessment, curation, augmentation, weak supervision, and governance for ML. https://arxiv.org/

Data Organization & Leakage Prevention

Oxford Flowers 102 Dataset

Preprocessing & Transform Pipelines

Data Augmentation & Robustness