Research map / Learning theory and algorithms / Prediction and decision losses
xRFM: Accurate, scalable, and interpretable feature learning models for tabular data
Daniel Beaglehole, David Holzmüller, Adityanarayanan Radhakrishnan, Mikhail Belkin
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 |
|---|---|---|---|
| add tree-based data partitioning with local leaf RFMs on top of original RFM (ablation) | xRFM · TALENT largest datasets (>=1 split) | mean normalized error: 0.0503 -> 0.0379 | Improved |
| use supervised AGOP split directions instead of unsupervised PCA splits (ablation) | xRFM · meta-test large regression | arithmetic mean error: 0.5638 -> 0.5503 | Improved |
| add temperature-tuned soft routing ensembling over leaf RFMs (ablation) | xRFM AGOP split · meta-test large regression | arithmetic mean error: 0.5514 -> 0.5503 | Improved |
| add temperature-tuned soft routing ensembling over leaf RFMs (ablation) | xRFM AGOP split · meta-test large regression | average eval time (s): 77 -> 187 | Hurt |
| add temperature-tuned soft routing ensembling over leaf RFMs (ablation) | xRFM AGOP split · meta-test large classification | arithmetic mean error: 0.1764 -> 0.1758 | Improved |
| use random-forest split instead of AGOP split (with temperature tuning) (ablation) | xRFM · meta-test large regression | arithmetic mean error: 0.5503 -> 0.5489 | No clear effect |
| use AGOP splits instead of RF/PCA splits for speed (ablation) | xRFM · meta-test large classification | average fit time (s): 6410 vs 6535 (PCA + TT) and 6510 (RF + TT) | Improved |
| use xRFM instead of GBDTs and TabPFN-v2 on TALENT regression | TALENT regression · TALENT regression (100 datasets) | aggregate RMSE | Improved |
| compare xRFM to kernel ridge regression and standard RFM (ablation) | xRFM · TALENT regression | normalized R2 | Improved |
| xRFM on classification vs best baselines | xRFM · TALENT classification (200 datasets) | rank | Mixed |
| xRFM vs original RFM on individual large datasets (Rain-in-Australia, diabetes-130-us-hospitals) (ablation) | xRFM · TALENT largest datasets | normalized error: 0.0535 -> 0.0629 (Rain-in-Australia) | Hurt |
| xRFM versus kernel solvers on large meta-test datasets | xRFM · meta-test 17 large datasets | percentage improvement over MLP | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- LightGBM — LightGBM: A Highly Efficient Gradient Boosting Decision Tree (NeurIPS 2017) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- A Data-Centric Perspective on Evaluating Machine Learning Models for Tabular Data (NeurIPS 2024 · similarity 0.88)
- Optimized Feature Generation for Tabular Data via LLMs with Decision Tree Reasoning (NeurIPS 2024 · similarity 0.88)
- TabPrep: Closing the Feature Engineering Gap in Tabular Benchmarks (NeurIPS 2026 · similarity 0.87)
- TabDPT: Scaling Tabular Foundation Models on Real Data (NeurIPS 2025 · similarity 0.87)
- TabFlex: Scaling Tabular Learning to Millions with Linear Attention (ICML 2025 · similarity 0.87)
- GRANDE: Gradient-Based Decision Tree Ensembles for Tabular Data (ICLR 2024 · similarity 0.87)
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.
