Identifying Representational Biases in Datasets Using PCA: A Max-Disparity Partition Framework
arXiv:2609.24556v1 Announce Type: new Abstract: Principal Component Analysis (PCA) minimises aggregate reconstruction error, which can inadvertently represent majority subgroups with substantially higher fidelity than minority subgroups. Fairness-aware extensions of PCA correct this disparity but r…
Read the full story at arXiv stat.ML ↗
Timeline · 1 report
- 2026-09-22 04:00 · arXiv stat.ML
Identifying Representational Biases in Datasets Using PCA: A Max-Disparity Partition Framework