Research map / Architectures and efficiency / Attention and transformers
TabSwift: An Efficient Tabular Foundation Model with Row-Wise Attention
Si-Yang Liu, Han-Jia Ye
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 learned exit head with register conditioning instead of entropy stopping at about 14-15 layers (ablation) | TabSwift early exit · TALENT classification | average accuracy: 0.8370 at 14.06 layers vs 0.8367 at 15.27 layers | Improved |
| use register-conditioned exit head instead of one without registers at tau 0.999 (ablation) | TabSwift early exit · TALENT classification | average accuracy: 0.8390 -> 0.8398 | Improved |
| early exit with threshold 0.5 instead of full depth | TabSwift · TALENT classification | average accuracy: 0.8398 -> 0.8334 | No clear effect |
| early exit with learned head w/ registers tau 0.9 instead of full depth | TabSwift · TALENT classification | average accuracy: 0.8398 -> 0.8329 | No clear effect |
| tighten learned exit threshold to 0.999 with registers | TabSwift early exit · TALENT classification | average accuracy: 0.8398 -> 0.8398 | No clear effect |
| evaluate on datasets with more than 20,000 training rows | TabSwift · TALENT classification | average rank: 12.32 vs RealMLP 9.31 | Hurt |
| compare with TabPFN v2 on high-dimensional datasets with PCA to 100 dims | TabSwift · High-dimensional datasets | average rank: 2.1 vs 1.8 | Mixed |
| TabSwift vs TabICL on high-dimensional datasets | TabSwift · High-dimensional datasets | average rank: 2.9 -> 2.1 | Improved |
Closest papers by meaning
- TabDPT: Scaling Tabular Foundation Models on Real Data (NeurIPS 2025 · similarity 0.92)
- Benchmarking Attention for Tabular Foundation Models (NeurIPS 2026 · similarity 0.92)
- TabICL: A Tabular Foundation Model for In-Context Learning on Large Data (ICML 2025 · similarity 0.91)
- TabPFN Unleashed: A Scalable and Effective Solution to Tabular Classification Problems (ICML 2025 · similarity 0.91)
- Understanding the Surprising Generalization Properties of Tabular Foundation Models (NeurIPS 2026 · similarity 0.90)
- Tabular Foundation Model for Generative Modelling (NeurIPS 2026 · similarity 0.90)
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.
