πŸ”¬

Robust Statistics

1
Open Unknowns
1
Cross-Domain Bridges
2
Active Hypotheses

Cross-Domain Bridges

Bridge RANSAC-style robust estimation and astronomical source matching share an outlier-dominated geometry problem: infer a transformation or correspondence from sparse inliers while cosmic rays, blends, artifacts, and catalog mismatches act as structured outliers.

Fields: Robust Statistics, Astronomy, Computer Science

The bridge is methodological. Astronomical cross-matching can use robust geometric-estimation ideas, but sky-survey outliers are not uniformly random, so standard RANSAC sampling assumptions require d...

Open Unknowns (1)

Unknown What is the optimal sampling strategy for RANSAC when the inlier distribution is non-uniform (spatially clustered, class-imbalanced, or structured) ΓÇâ and can information-theoretic bounds on robust estimation be derived for heterogeneous inlier distributions? u-ransac-optimal-sampling-strategy-non-uniform-inlier-distribution

Active Hypotheses

Hypothesis In simulated astronomical cross-matching with clustered artifacts, quality-ranked RANSAC sampling will reduce false matches by at least 20 percent versus uniform RANSAC at fixed recall; falsified if ranked sampling overfits survey-quality flags and loses recall. medium
Hypothesis Applying robust statistics estimators (Huber loss, trimmed loss with formal breakdown point guarantees) to deep neural network training with noisy labels will outperform standard cross-entropy training and ad hoc noise-robust methods when label noise exceeds 20% ΓÇâ because the formal 50% breakdown point provides a principled bound that heuristic methods lack. medium

Know something about Robust Statistics? Contribute an unknown or hypothesis β†’

Generated 2026-05-10 Β· USDR Dashboard