Identifiable but Hard to Estimate: Rotation in Equally Spaced Overcomplete ICA
Organizations: The George Washington University · University of Health and Rehabilitation Sciences · University of California, San Diego · Georgia State University ·
Abstract
How many observations are needed to estimate a common rotation in two-dimensional overcomplete independent component analysis? We study equally spaced mixing columns with known source count, common source law, and column norm. For a fixed non-Gaussian fourth-order Gram-Charlier source law, the rotation is identifiable modulo the column spacing. Yet the sample size for fixed relative angular accuracy grows exponentially with the source count. We characterize this sample size, up to absolute constants, by the reciprocal of the squared Hellinger distance between two rotations separated by half the column spacing. An explicit radially weighted phase estimator attains this scale uniformly over the rotation family, using the observation radius to prevent cancellation in the mean of an angular harmonic. We also determine the exact exponential decay rate of single-observation total variation for the same pair. This gives bounds on the sample size exponent; its exact value remains open. Finally, a noisy binary-source example shows how an exact zero line of the observed characteristic function can have an exponentially small normal derivative.
Figures & tables
Appendix figures & tables5 assets
Supplementary material from the paper’s appendix.
Appendix
| Max. log residual | Theoretical upper bound (numerically evaluated) | ||||
|---|---|---|---|---|---|
| 1 | 24 | 3.91120 | 0.4940 | 3.91128 | |
| 1 | 30 | 3.91121 | 0.4946 | 3.91128 | |
| 4 | 24 | 2.36864 | 0.4864 | 2.3689 | |
| 4 | 30 | 2.36873 | 0.4900 | 2.3689 |
| Refined | Relative change | |
|---|---|---|
| 10 | No change at stored precision | |
| 56 |
| Analysis | Source family and supplied information | Target | Scope and limit |
|---|---|---|---|
| Theorems 1 – 2 | Gram–Charlier; , source law, spacing and column norm known | Common rotation modulo | Global, finite ; exponents at fixed , |
| Appendix B | Independent sources; known target real zero; fixed or independent isolating window; unique grid minimum | Labeled direction and zero-line offset | Local; fixed model and grid, |
| Appendix C | Two sources under Assumption A; target label known; source laws, scales and second direction unknown | Labeled direction | Local risk; fixed finite jump paths before , |
| Appendix D | Known Gram–Charlier law and equal geometry; one complex zero and an isolating angle interval | Common rotation near the true angle | Local; first at fixed , then variance coefficient as |
Independent Component Analysis
Independent component analysis separates observed mixtures into statistically independent latent signals for source recovery and data analysis.
See all 24 papers