OpenML Benchmark
Cross-source consensus on OpenML Benchmark from 1 sources and 5 claims.
1 sources · 5 claims
How it works
Benefits
Preparation
Comparisons
Highlighted claims
- The OpenML benchmark used NSGA-II through Optuna with objectives to maximize kNN accuracy and minimize topology error. — DiRe-RAPIDS: Topology-faithful dimensionality reduction at scale
- Tuned DiRe achieved topology error 0 on all 11 OpenML datasets. — DiRe-RAPIDS: Topology-faithful dimensionality reduction at scale
- At least one Pareto-optimal DiRe configuration dominated cuML UMAP on both kNN accuracy and topology error in 7 of 11 datasets. — DiRe-RAPIDS: Topology-faithful dimensionality reduction at scale
- In the covertype case, all three Pareto-optimal DiRe trials strictly dominated cuML UMAP. — DiRe-RAPIDS: Topology-faithful dimensionality reduction at scale
- The Pareto-optimal hyperparameters converged on spectral initialization, larger spread, moderate-to-high iterations, and 10 to 30 neighbors. — DiRe-RAPIDS: Topology-faithful dimensionality reduction at scale