English

Constraint-Driven Warm-Freeze for Efficient Transfer Learning in Photovoltaic Systems

Neural and Evolutionary Computing 2026-04-08 v1

Abstract

Detecting cyberattacks in photovoltaic (PV) monitoring and MPPT control signals requires models that are robust to bias, drift, and transient spikes, yet lightweight enough for resource-constrained edge controllers. While deep learning outperforms traditional physics-based diagnostics and handcrafted features, standard fine-tuning is computationally prohibitive for edge devices. Furthermore, existing Parameter-Efficient Fine-Tuning (PEFT) methods typically apply uniform adaptation or rely on expensive architectural searches, lacking the flexibility to adhere to strict hardware budgets. To bridge this gap, we propose Constraint-Driven Warm-Freeze (CDWF), a budget-aware adaptation framework. CDWF leverages a brief warm-start phase to quantify gradient-based block importance, then solves a constrained optimization problem to dynamically allocate full trainability to high-impact blocks while efficiently adapting the remaining blocks via Low-Rank Adaptation (LoRA). We evaluate CDWF on standard vision benchmarks (CIFAR-10/100) and a novel PV cyberattack dataset, transferring from bias pretraining to drift and spike detection. The experiments demonstrate that CDWF retains 90 to 99% of full fine-tuning performance while reducing trainable parameters by up to 120x. These results establish CDWF as an effective, importance-guided solution for reliable transfer learning under tight edge constraints.

Keywords

Cite

@article{arxiv.2604.05807,
  title  = {Constraint-Driven Warm-Freeze for Efficient Transfer Learning in Photovoltaic Systems},
  author = {Yasmeen Saeed and Ahmed Sharshar and Mohsen Guizani},
  journal= {arXiv preprint arXiv:2604.05807},
  year   = {2026}
}

Comments

Preprint submitted and accepted to IEEE IJCNN (WCCI)

R2 v1 2026-07-01T11:57:18.460Z