EUVIMEDEUVIMED entdecken
Open Access · CC BY-SA 4.0

HMCO-AT: A Hierarchical Memory - Compute Orchestrated Adaptive Training Framework for Resource-Efficient and Adversarially Robust Large Models in Cloud–Edge Systems

Yutong Zhou, Liesheng Wei

Journal of Computer Science and Frontier Technologies · 16.06.2026 · 3 (3) · 37

DOI: 10.63313/jcsft.9085 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Computer Science and Frontier Technologies, online erschienen am 16.06.2026. Verfasst von Yutong Zhou und 1 weiteren Autor:innen. Die Arbeit nennt 29 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-SA 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Adversarial training has emerged as the most effective paradigm for improving the robustness of deep neural networks; however, it imposes substantial computational and memory demands, particularly during the adversarial example generation process. These challenges become more critical in cloud-edge collaborative learning scenarios, where edge devices have limited resources and communication bandwidth is highly variable. To address this problem, we propose HMCO-AT, a Hierarchical Memory-Compute Orchestrated Adaptive Training framework designed specifically for resource-efficient adversarial training of large-scale models. HMCO-AT explicitly models the memory footprint and computation peaks induced by adversarial sample generation and dynamically schedules model layers, training stages, and communication operations across cloud and edge nodes. Our framework integrates hierarchical parameter partitioning, memory-aware activation recomputation, quantized gradient transmission, and an RL-based adaptive scheduler that jointly optimizes memory consumption, computational load, and robustness performance. HMCO-AT is validated on large-scale language-modeling corpora including OpenWebText, WikiCorpus, and filtered C4 under realistic cloud–edge conditions. Experiments show that HMCO-AT reduces peak edge memory by 64.8% and lowers computation cost and latency by 35.8% and 26.7%, respectively. It also improves adversarial robustness, achieving 7.2% higher FGSM and 9.3% higher PGD-10 accuracy than strong baselines. These results confirm HMCO-AT as an efficient and robust framework for large-model cloud–edge training.

Zitieren

APA

Zhou, Y., & Wei, L. (2026). HMCO-AT: A Hierarchical Memory - Compute Orchestrated Adaptive Training Framework for Resource-Efficient and Adversarially Robust Large Models in Cloud–Edge Systems. Journal of Computer Science and Frontier Technologies, 3(3), 37. https://doi.org/10.63313/jcsft.9085

Vancouver

Zhou Y, Wei L. HMCO-AT: A Hierarchical Memory - Compute Orchestrated Adaptive Training Framework for Resource-Efficient and Adversarially Robust Large Models in Cloud–Edge Systems. Journal of Computer Science and Frontier Technologies. 2026;3(3):37. doi:10.63313/jcsft.9085

BibTeX
@article{Zhou2026,
  author = {Zhou, Yutong and Wei, Liesheng},
  title = {HMCO-AT: A Hierarchical Memory - Compute Orchestrated Adaptive Training Framework for Resource-Efficient and Adversarially Robust Large Models in Cloud–Edge Systems},
  journal = {Journal of Computer Science and Frontier Technologies},
  year = {2026},
  volume = {3},
  number = {3},
  pages = {37},
  doi = {10.63313/jcsft.9085}
}
Kostenlos starten →