@article{Lun2026, 
author = {Darren Chai Xin Lun and Lim Tong Ming},
title = {H-LoRA: Rethinking Rank Selection for Controllable Knowledge Retention in Edge AI},
year = {2026},
journal = {Computers, Materials & Continua},
volume = {88},
number = {1},
keywords = {LoRA, edge AI, knowledge retention, domain adaptation, parameter-efficient fine-tuning, catastrophic forgetting},
url = {https://www.sciopen.com/article/10.32604/cmc.2026.080068},
doi = {10.32604/cmc.2026.080068},
abstract = {The deployment of specialized language models in resource-constrained edge environments ( ≤1B parameters,  ≤2 GB memory,  ≤100 ms latency) faces a critical challenge: Supervised Fine-Tuning (SFT) achieves domain expertise but suffers from irreversible catastrophic forgetting, while traditional Low-Rank Adaptation (LoRA) with conservative ranks ( r ≤ 64) often underperforms due to insufficient adaptation capacity. This work introduces H-LoRA (High-Rank LoRA) for edge-deployable models and establishes a fundamental distinction between destructive forgetting and controllable knowledge retention. Through comprehensive experiments on compact models ( 0.12B Minimind and Qwen- 0.5B) across three domains (Human Resources, Medical, Mathematics) using 29,647 samples, we demonstrate that while both SFT and H-LoRA exhibit general capability degradation, they differ fundamentally: SFT completely destroys the original knowledge structure ( 1% topic retention), while H-LoRA maintains knowledge integrity with  90% topic retention—an  89 percentage point improvement—enabling post-deployment capability recovery. H-LoRA employs simplified scaling and strategic high-rank adaptation at approximately two-thirds of the model’s hidden dimension ( r = 512 for  d = 768), achieving SFT-level domain performance ( 99.81% precision) with  5× greater parameter efficiency ( 20.35% trainable parameters) and robust cross-domain generalization ( 93.5 ± 6.8% average precision). In addition, H-LoRA reduces over-the-air (OTA) update size from  1.4 GB to  96 MB ( ≈93%), enabling practical and frequent deployment of specialized models in bandwidth-limited edge environments. Beyond demonstrating effectiveness, this work establishes the first comprehensive framework for characterizing specialization-retention trade-offs in parameter-efficient fine-tuning, providing practical guidance for method selection in real-world deployments.}
}