Paper Overview
Field: Machine Learning Authors: Yury Gorishniy, Ivan Rubachev, Dmitrii Feoktistov Published: 2025-04-17 arXiv: 2504.13081
Summary
MLP is a heavily used backbone in modern deep learning architectures for supervised learning on tabular data, and AdamW is the go-to optimizer used to train tabular DL models. Unlike architecture design, however, the choice of optimizer for tabular DL has not been examined systematically, despite new optimizers showing promise in other domains. To fill this gap, the authors benchmark N optimizers on N tabular datasets for training MLP-based models in the standard supervised learning setting under a shared experiment protocol.
Key Findings
- Muon outperforms AdamW: The Muon optimizer consistently outperforms AdamW, and thus should be considered a strong and practical choice for practitioners and researchers, if the associated training efficiency overhead is affordable.
- EMA helps AdamW on vanilla MLPs: Exponential moving averaging of model weights is a simple yet effective technique that improves AdamW's performance on vanilla MLPs, though its effect is less consistent across other model variants.
Original Abstract (excerpt)
MLP is a heavily used backbone in modern deep learning (DL) architectures for supervised learning on tabular data, and AdamW is the go-to optimizer used to train tabular DL models. Unlike architecture design, however, the choice of optimizer for tabular DL has not been examined systematically, despite new optimizers showing promise in other domains. To fill this gap, we benchmark N optimizers on N tabular datasets for training MLP-based models in the standard supervised learning setting under a shared experiment protocol.
Read the full paper: https://arxiv.org/abs/2504.13081