Research map / Learning theory and algorithms / Prediction and decision losses
High Dimensional Mode Hunting Using Pettiest Components Analysis
Tianhao Liu, Daniel Andrés Díaz Pachón, J. Sunil Rao, Jean-Eudes Dazard
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 |
|---|---|---|---|
| project onto 2 pettiest instead of 2 principal components (ablation) | fastPRIM · Simulation (p=100, beta=0.1) | box density per volume (step 1): 16.3 -> 215 | Improved |
| project onto 2 pettiest instead of 2 principal components (ablation) | PRIM · Simulation (p=100, beta=0.1) | box density per volume (step 1): 15.7 -> 182 | Improved |
| reduce dimension 100 to 2 with principal components (ablation) | PRIM · Simulation | box density per volume (step 1): 1.07e-73 -> 15.7 | Improved |
| use fastPRIM instead of PRIM (ablation) | pettiest components · Simulation | box density per volume (step 1): 182 -> 215 | Improved |
| project onto pettiest instead of principal components (ablation) | PRIM · Simulation | empirical variance: 2.28 -> 0.204 | Improved |
| project onto pettiest instead of principal components (ablation) | fastPRIM · Simulation | empirical variance: 2.23 -> 0.149 | Improved |
| use fastPRIM instead of PRIM (ablation) | principal components · Simulation | empirical bias: 0.145 -> 0.00579 | Improved |
| use fastPRIM instead of PRIM (ablation) | pettiest components · Simulation | empirical bias: 0.1 -> 0.000653 | Improved |
| project onto pettiest instead of principal components (ablation) | fastPRIM · MNIST digit 1 | active information (bits): -1.21 -> 7.96 | Improved |
| project onto pettiest instead of principal components (ablation) | fastPRIM · MNIST digit 8 | active information (bits): 1.32 -> 3.36 | Improved |
| project onto pettiest instead of principal components (ablation) | PRIM · MNIST digit 0 | active information (bits): 1.29 -> 6.26 | Improved |
| use fastPRIM instead of PRIM with pettiest components (ablation) | PRIM-Pettiest · MNIST digits 0, 1 and 5 | active information (bits): 8.08 -> 7.96 (digit 1) | Hurt |
Closest papers by meaning
- Solving Large-Scale Sparse PCA to Certifiable (Near) Optimality (JMLR 2022 · similarity 0.82)
- Disentangling the Modes of Variation in Unlabelled Data (TPAMI 2017 · similarity 0.82)
- Sparse PCA: a Geometric Approach (JMLR 2023 · similarity 0.82)
- Principle Component Trees and Their Persistent Homology (AAAI 2024 · similarity 0.82)
- Capturing the denoising effect of PCA via compression ratio (NeurIPS 2024 · similarity 0.81)
- Efficient Sparse PCA via Block-Diagonalization (ICLR 2025 · similarity 0.81)
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.
