Research map / Learning theory and algorithms / Matrix and tensor methods
Multivariate Regression with Gross Errors on Manifold-Valued Data
Xiaowei Zhang, Xudong Shi, Yu Sun, Li Cheng
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| model gross error explicitly (PALMR instead of MGLM) | geodesic regression · C-MIND DTI slice z=32, 20% manual gross error | MSGE: 0.3530 -> 0.2196 | Improved |
| model gross error explicitly (PALMR instead of MGLM) | geodesic regression · C-MIND DTI slice z=24, 20% manual gross error | relative FA error: 0.7342 -> 0.5081 | Improved |
| model gross error explicitly (PALMR instead of MGLM) | geodesic regression · C-MIND DTI slice z=32, 20% registration error | relative FA error: 0.1654 -> 0.1316 | Improved |
| model gross error explicitly (PALMR instead of MGLM) | geodesic regression · C-MIND DTI slice x=64, 20% registration error | MSGE: 0.2436 -> 0.2226 | Improved |
| model gross error explicitly (PALMR instead of MGLM) | geodesic regression · C-MIND DTI slice z=32, no gross error | relative FA error: 0.3223 -> 0.3210 | No clear effect |
| model gross error explicitly (PALMR instead of MGLM) | geodesic regression · C-MIND DTI slice y=64, no gross error | relative FA error: 0.3162 -> 0.3166 | No clear effect |
| regress on full tensors with geodesic regression instead of FA values | PALMR/MGLM vs FA regression · C-MIND DTI slice z=32, no gross error | relative FA error: 0.9376 -> 0.3210 | Improved |
| use PALMR instead of best competitor under manual and registration gross error | PALMR · C-MIND DTI six slices | median prediction error improvement: at least 20% and 15% | Improved |
| use PALMR instead of MGLM when more than 80% of training data corrupted | PALMR · synthetic DTI, varying beta | MSGE: performance slightly worse above 80% corruption | Hurt |
| use manifold-aware PALMR instead of Euclidean robust regression on vectorized tensors | PALMR · synthetic DTI, varying gross error magnitude and ratio | Frobenius error | Improved |
| use MGLM (no gross error model) instead of Euclidean model under large gross errors | MGLM · synthetic DTI, large sigma_g | Frobenius error | Mixed |
Closest papers by meaning
- A Nonlinear Regression Technique for Manifold Valued Data With Applications to Medical Image Analysis (CVPR 2016 · similarity 0.89)
- RMLR: Extending Multinomial Logistic Regression into General Geometries (NeurIPS 2024 · similarity 0.85)
- Diffusion Models and the Manifold Hypothesis: Log-Domain Smoothing is Geometry Adaptive (NeurIPS 2025 · similarity 0.85)
- The Noisy Laplacian: a Threshold Phenomenon for Non-Linear Dimension Reduction (ICML 2025 · similarity 0.85)
- High-Dimensional Tensor Regression With Oracle Properties (ICML 2025 · similarity 0.84)
- Intrinsic Riemannian Cross-covariance for Manifold-valued Random Objects (NeurIPS 2026 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
