Laws of Learning Dynamics and the Core of Learners
Fuente:
arXiv
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917248888733696 |
|---|---|
| author | Jung, Inkee Lau, Siu Cheong |
| author_facet | Jung, Inkee Lau, Siu Cheong |
| contents | We formulate the fundamental laws governing learning dynamics, namely the conservation law and the decrease of total entropy. Within this framework, we introduce an entropy-based lifelong ensemble learning method. We evaluate its effectiveness by constructing an immunization mechanism to defend against transfer-based adversarial attacks on the CIFAR-10 dataset. Compared with a naive ensemble formed by simply averaging models specialized on clean and adversarial samples, the resulting logifold achieves higher accuracy in most test cases, with particularly large gains under strong perturbations. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_05026 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Laws of Learning Dynamics and the Core of Learners Jung, Inkee Lau, Siu Cheong Machine Learning Artificial Intelligence Information Theory 94A17, 62B10, 68Q32, 68T05 I.2.6; H.1.1; I.2.0 We formulate the fundamental laws governing learning dynamics, namely the conservation law and the decrease of total entropy. Within this framework, we introduce an entropy-based lifelong ensemble learning method. We evaluate its effectiveness by constructing an immunization mechanism to defend against transfer-based adversarial attacks on the CIFAR-10 dataset. Compared with a naive ensemble formed by simply averaging models specialized on clean and adversarial samples, the resulting logifold achieves higher accuracy in most test cases, with particularly large gains under strong perturbations. |
| title | Laws of Learning Dynamics and the Core of Learners |
| topic | Machine Learning Artificial Intelligence Information Theory 94A17, 62B10, 68Q32, 68T05 I.2.6; H.1.1; I.2.0 |
| url | https://arxiv.org/abs/2602.05026 |