Jingya Zheng

Jingya Zheng

Auto-generated profile, based on publicly available information

Jingya Zheng

Shandong Future Network Research Institute, Jinan, China

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
23 April 2025 in IEEE Access
Verified
Journal Article
Open access
JZJingya Zheng
Jingya Zheng
GTGaofeng Tao
Gaofeng Tao
SQShuxin Qin
Shuxin Qin
DWDan Wang
Dan Wang
ZMZhongjun Ma
Zhongjun Ma
Storage resources are essential in heterogeneous multi-cloud environments. In response to the growing demand for efficient storage resource management (SRM) in these environments, this paper proposes an intent-based storage management (IBSM) system powered by a fine-tuned, lightweight large language model (LLM). To overcome the limitations of existing methods, the IBSM system focuses on enhancing the controllability, completeness, and reliability of SRM in multi-cloud environments. Specifically, the IBSM system employs a dual-phase joint intent classification algorithm, which leverages a fine-tuned LLM to accurately identify user intents across diverse knowledge backgrounds. Additionally, the system constructs a collaborative intent decomposition method, which guarantees the integrity of intents. Furthermore, the system integrates an automated intent deployment mechanism that supports error recovery through checkpoints. Experimental results show that the system achieves a whole end-to-end (E2E) lifecycle for managing user intents. The E2E time is reduced by at least half compared to the manual approach, with an average of 50.14% dedicated to interactive tasks. Performance metrics for intent classification, including accuracy, precision, and recall, all exceed 90%. Moreover, the recovery time is reduced by an average of 30.6%. Therefore, the system provides a valuable solution for the autonomous management of multi-cloud resources.

17 December 2024 in IEEE Transactions on Computational Social Systems
Verified
Journal Article
Closed access
SQShuxin Qin
Shuxin Qin
YLYongcan Luo
Yongcan Luo
JZJing Zhu
Jing Zhu
GTGaofeng Tao
Gaofeng Tao
JZJingya Zheng
Jingya Zheng
ZMZhongjun Ma
Zhongjun Ma
Detecting abnormal nodes from attributed networks plays an important role in various applications, including cybersecurity, finance, and social networks. Most existing methods focus on learning different scales of graphs or using augmented data to improve the quality of feature representation. However, the performance is limited due to two critical problems. First, the high sensitivity of attributed networks makes it uncontrollable and uncertain to use conventional methods for data augmentation, leading to limited improvement in representation and generalization capabilities. Second, under the unsupervised paradigm, anomalous nodes mixed in the training data may interfere with the learning of normal patterns and weaken the discrimination ability. In this work, we propose a novel multiview and multiscale contrastive learning framework to address these two issues. Specifically, a network augmentation method based on parameter perturbation is introduced to generate augmented views for both node–node and node–subgraph level contrast branches. Then, cross-view graph contrastive learning is employed to improve the representation without the need for augmented data. We also provide a cycle training strategy where normal samples detected in the former step are collected for an additional training step. In this way, the ability to learn normal patterns is enhanced. Extensive experiments on six benchmark datasets demonstrate that our method outperforms the existing state-of- the-art baselines.