APAR: Modeling Irregular Target Functions in Tabular Regression via Arithmetic-Aware Pre-Training and Adaptive-Regularized Fine-Tuning
Hong-Wei Wu, Wei-Yao Wang, Kuang-Da Wang, Wen-Chih Peng · 2024
Read APAR: Modeling Irregular Target Functions in Tabular Regression via Arithmetic-Aware Pre-Training and Adaptive-Regularized Fine-Tuning by Hong-Wei Wu, Wei-Yao Wang, Kuang-Da Wang, Wen-Chih Peng (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-10941.txt.
Description
Hong-Wei Wu and co-authors present APAR in 2024, a framework for predicting values from table data. Tables drive work from finance to health, yet deep networks often fail to beat older tree methods there. The snag is jagged target functions, where small input shifts swing the answer.
Pre-training teaches the model arithmetic-aware links between rows and their continuous labels. Fine-tuning then adds a self-taught consistency rule that curbs overfitting. Across ten data sets the method beat tree, network, and hybrid rivals on error scores. Ablation tests confirmed each stage helped. Code and data are public for others to build on.
Catalog record
| CREATOR | Hong-Wei Wu, Wei-Yao Wang, Kuang-Da Wang, Wen-Chih Peng |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 48 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-10941