Abstract:We propose Evolutionary Dynamic Loss (EDL), a framework that learns a transferable classification loss in the probability space using unlimited synthetic prediction-label pairs, without accessing real samples during the main loss pretraining stage. EDL parameterizes the loss as a lightweight network and is trained with a semantics-free ranking-consistency objective that assigns larger penalties for more erroneous predictions. To robustly explore the space of loss functions, we optimize EDL via an evolutionary strategy and introduce chaotic mutation to improve exploration under noisy fitness evaluations. Experiments on CIFAR-10 with ResNet backbones show that EDL can serve as a drop-in replacement for cross-entropy and achieves competitive or improved accuracy, while ablation studies confirm that chaotic mutation yields faster convergence and better synthetic pretraining metrics than standard Gaussian mutation.
| Comments: | 6 pages |
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2605.03722 [cs.LG] |
| (or arXiv:2605.03722v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2605.03722 arXiv-issued DOI via DataCite (pending registration) |
From: Xiang Meng [view email]
[v1]
Tue, 5 May 2026 13:08:18 UTC (4,149 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。