Research map / Label-efficient and robust learning / Recognition and long-tail
Scalable Feature Selection via Distributed Diversity Maximization
Sepehr Zadeh, Mehrdad Ghadiri, Vahab Mirrokni, Morteza Zadimoghaddam
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 diversity-based objective with distributed greedy (DDISMI) vs mRMR | DDISMI · NCI9 | CV accuracy (%) SVM: 74.4 -> 82.2 | Improved |
| use DDISMI vs mRMR | DDISMI · TOX 171 | CV accuracy (%) SVM: 83.8 -> 88.8 | Improved |
| use DDISMI vs mRMR | DDISMI · Leukemia | CV accuracy (%) SVM: 99.8 -> 96.1 | Hurt |
| distribute features across machines (DDISMI) instead of centralized GREEDY (ablation) | DDISMI · TOX 171 | CV accuracy (%) SVM: 88.3 -> 88.8 | No clear effect |
| distribute features across machines (DDISMI) instead of centralized GREEDY (ablation) | DDISMI · NCI9 | CV accuracy (%) 3-NN: 81.5 -> 80.1 | No clear effect |
| distribute features across machines (DDISMI) instead of GREEDY (ablation) | DDISMI · Binary news20, 10 features | running time speed-up: 164.1 | Improved |
| combine VI redundancy term and MI relevance term (lambda 0.8) versus either alone (ablation) | DDISMI · TOX 171 | classification accuracy | Improved |
| use greedy instead of local search for distributed diversity maximization (ablation) | DDISMI · NCI9 | objective value, accuracy, time | Improved |
| tune lambda per dataset (0.4 instead of 0.8) (ablation) | DDISMI · Leukemia | accuracy: 96.1 -> 99.9 | Improved |
Closest papers by meaning
- LESS-VFL: Communication-Efficient Feature Selection for Vertical Federated Learning (ICML 2023 · similarity 0.86)
- Greedy Column Subset Selection: New Bounds and Distributed Algorithms (ICML 2016 · similarity 0.86)
- Diversified Batch Selection for Training Acceleration (ICML 2024 · similarity 0.86)
- Rates of Convergence for Large-scale Nearest Neighbor Classification (NeurIPS 2019 · similarity 0.86)
- SAND: One-Shot Feature Selection with Additive Noise Distortion (ICML 2025 · similarity 0.86)
- GIST: Greedy Independent Set Thresholding for Max-Min Diversification with Submodular Utility (NeurIPS 2025 · 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.
