KAE: Kolmogorov-Arnold Auto-Encoder for Representation Learning
Fangchen Yu, Ruilizhen Hu, Yidong Lin, Yuqi Ma, Zhenghao Huang, Wenye Li · 2025
Read KAE: Kolmogorov-Arnold Auto-Encoder for Representation Learning by Fangchen Yu, Ruilizhen Hu, Yidong Lin, Yuqi Ma, Zhenghao Huang, Wenye Li (2025) 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-2501-00420.txt.
Description
Fangchen Yu and co-authors introduce in 2024 an autoencoder built from a new kind of network layer, aimed at learning better data representations. The layer type puts learnable curve functions on the links between units, which can raise accuracy and clarity against classic designs. Their model folds that layer into the standard squeeze and rebuild design of an autoencoder.
Flexible curve shapes let the model catch tangled patterns and curved links in data. On standard data sets, it beat plain autoencoders and sibling variants on search, sorting into classes, and cleaning noisy inputs, with lower rebuild error. Code is shared. The authors offer the model as a general tool for representation learning.
Catalog record
| CREATOR | Fangchen Yu, Ruilizhen Hu, Yidong Lin, Yuqi Ma, Zhenghao Huang, Wenye Li |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 36 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-00420