Research map / Learning theory and algorithms / Matrix and tensor methods
Online Dictionary Learning for Approximate Archetypal Analysis
Jieru Mei, Chunyu Wang, Wenjun Zeng
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 |
|---|---|---|---|
| use archetypes learned by the proposed method instead of sparse-coding bases | PCA-style 3D pose fitting · H3.6M (known camera) | MPJPE (mm): 231.82 -> 50.71 | Improved |
| use archetypes learned by the proposed method instead of sparse-coding bases | 3D pose fitting · H3.6M (estimated camera) | MPJPE (mm): 252.86 -> 76.91 | Improved |
| use sparse-coding bases instead of archetypes | 3D pose fitting · H3.6M (known camera) | 2D projection error (pixel): 3.68 -> 0.37 | Improved |
| use online approximate AA instead of original archetypal analysis | 3D pose fitting · H3.6M (known camera) | MPJPE (mm): 52.56 -> 50.71 | No clear effect |
| use online approximate AA instead of original archetypal analysis | 3D pose fitting · H3.6M (estimated 2D pose) | MPJPE (mm): 126.88 -> 106.52 | Improved |
| use online approximate AA instead of original archetypal analysis | 3D pose fitting · H3.6M (estimated camera, GT 2D) | MPJPE (mm): 75.93 -> 76.91 | No clear effect |
| replace sparse coding with archetypes | nearest-reconstruction classifier · NTU RGB+D Cross-Subject | accuracy: 64.95 -> 68.07 | Improved |
| replace sparse coding with archetypes | nearest-reconstruction classifier · NTU RGB+D Cross-View | accuracy: 77.61 -> 80.93 | Improved |
| compare with VA-LSTM deep model | archetype classifier · NTU RGB+D Cross-Subject | accuracy: 79.40 -> 68.07 | Hurt |
| use online dictionary updates instead of full-batch archetypal analysis | approximate AA · H3.6M, 400 archetypes | training speed | Improved |
Closest papers by meaning
- A Geometric Approach to Archetypal Analysis via Sparse Projections (ICML 2020 · similarity 0.90)
- A Survey on Archetypal Analysis (TPAMI 2026 · similarity 0.89)
- Coresets for Archetypal Analysis (NeurIPS 2019 · similarity 0.89)
- Archetypal SAE: Adaptive and Stable Dictionary Learning for Concept Extraction in Large Vision Models (ICML 2025 · similarity 0.86)
- Bridging Neural and Symbolic Representations with Transitional Dictionary Learning (ICLR 2024 · similarity 0.81)
- Navigating the Effect of Parametrization for Dimensionality Reduction (NeurIPS 2024 · similarity 0.80)
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.
