Diming Zhang

Prof. Dr. Diming Zhang


Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

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

Representative Publications

View All Publications
1 July 2025 in Scientific Reports
Verified
Journal Article
Open access
PDF file
DZDiming Zhang
Diming Zhang
YXYuchen Xu
Yuchen Xu
YLYuanjiang Li
Yuanjiang Li
Electric load forecasting’s accuracy and reliability are pivotal for enhancing the dispatch efficiency of power systems and the integration of renewable energy into the grid. In response to this need, this paper introduces a novel electric load forecasting framework. Initially, to address the limitations of the Sparrow Search Algorithm (SSA), we propose a multi-strategy improved SSA with dynamic inertia weight (WHFSSA). This enhancement involves balancing the global and local search capabilities by introducing dynamic inertia weights, employing a suboptimal solution guidance strategy to facilitate the escape from local optima, and applying dimension-wise dynamic reverse learning to boost population diversity and quality, thereby hastening convergence. Subsequently, we optimize the parameters of Variational Mode Decomposition (VMD) using the improved SSA (WHFSSA) to achieve precise decomposition of electric load sequences into more regular subsequences. These subsequences are then integrated with the Kernel Extreme Learning Machine (KELM) to develop a forecasting model named WHFSSA-KELM. The efficacy of this model is validated on two electric load datasets across different forecasting tasks. In Dataset 1, the focus is on the impact of temperature and historical load on future load values. The results show that the proposed model realizes an average improvement of 5.7% in the R2 metric compared to benchmark models. In Dataset 2, which considers various additional factors influencing electric load, the model’s performance is assessed for three-step-ahead forecasting. It achieves average R2 metric improvements of 5.6%, 7.6%, and 10.9% for the three-step-ahead forecasts, respectively, compared to benchmark models. Consequently, the proposed method offers more accurate forecasting, contributing to the safe and stable operation of power systems.

28 April 2026 in Sensors
Verified
Journal Article
Open access
PDF file
XKXu Kai
Xu Kai
DZDiming Zhang
Diming Zhang
XWXuguo Wang
Xuguo Wang
To address the bottlenecks of missing decision-making closed loop, insufficient experience reuse, and decoupled resource scheduling in industrial LLM deployment, this paper proposes LLM-Conductor, a three-layer collaborative architecture that enables monitoring-feedback autonomous decision-making, structured policy memory, and joint policy-resource optimization.Through ablation studies, horizontal comparisons with ISOLATEGPT and ReAct, and graded resource-reduction experiments across six tiers, the results demonstrate that the security risk incidence rate is reduced from 70.6 percent to 1.3 percent, the multi-application collaborative task completion rate reaches 100 percent, and token utilization improves to 88.9 percent. Under constraints of at least 512 MB memory and at least 0.5 GHz CPU, the core task completion rate remains above 95 percent. By deeply coupling decision-making with resource scheduling, this architecture provides an integrated pathway toward efficient, secure, and reliable LLM deployment in Industrial Internet of Things scenarios. Current validation focuses on software-layer interaction patterns under simulated resource-constrained environments, with physical-layer industrial integration reserved for future work.

14 August 2026 in Measurement Science and Technology
Verified
Journal Article
Closed access
HGHaoyu Gu
Haoyu Gu
DZDiming Zhang
Diming Zhang
SYShichao Yi
Shichao Yi
CWChengyi Wang
Chengyi Wang
In industrial internet of things (IIoT) deployments, condition monitoring data is rarely complete. Labeled samples are scarce, sensors drop out, signals are corrupted by noise, and what reaches the edge is often a sparse, distorted fragment of the ideal dataset. Federated learning preserves privacy, yet its standard formulations assume clients can train locally on adequate labeled data-an assumption that seldom holds on the factory floor. This paper proposes prompt transfer, a distinct federated learning paradigm that replaces conventional gradient or prototype exchange with the transfer of small, curated sets of labeled examples (prompts) as the sole medium of knowledge flow. Unlike federated distillation, which shares model outputs, or meta-learning, which requires local gradient adaptation, prompt transfer leverages the in-context learning capability of a frozen tabular foundation model to achieve zero-gradient, parameter-free knowledge transfer across clients. Experiments on the Southeast University bearing dataset show that under 5-shot settings, our method FedICL achieves 94.0% accuracy, in contrast to FedAvg’s random-guessing performance of 20.1% while requiring zero gradient computation and reducing communication rounds by 95% and data transfer by 87%. The approach remains effective under non-IID distributions ( 53.9 % vs 20 % random). This work directly addresses the special issue themes of incomplete data and federated learning, offering a practical solution for label-scarce, resource-constrained IIoT scenarios.

15 May 2026 in Sensors
Verified
Journal Article
Open access
PDF file
YGYuchang Gu
Yuchang Gu
DZDiming Zhang
Diming Zhang
TLTaiyu Lu
Taiyu Lu
Dynamic scheduling can improve the latency and memory efficiency of deep neural network inference on edge devices, but it often introduces cold-start overhead when a newly deployed model requires online profiling and policy adaptation before reaching stable performance. This paper proposes EdgeOpt-Sched-CS, a cold-start-aware extension of dynamic graph scheduling for edge inference. The key idea is to initialize the scheduler of a target computation graph using scheduling knowledge transferred from structurally similar source graphs, instead of starting from a generic policy. EdgeOpt-Sched-CS constructs compact graph signatures, retrieves relevant source schedulers, and performs lightweight cold-start-aware online adaptation during early deployment. We evaluate the framework across representative device–model scenarios involving lightweight convolutional neural networks, transformer models, and quantized language-model workloads. The results show that EdgeOpt-Sched-CS reduces cumulative cold-start latency by 10.6–20.4% and shortens time-to-stability by 5.2–21.7%, while preserving the steady-state latency–memory behavior of the original dynamic scheduler with only small additional scheduling overhead. These findings indicate that scheduler initialization is an important optimization dimension for adaptive edge inference and that prior scheduling knowledge can be effectively reused across related computation graphs.

18 May 2026 in Measurement Science and Technology
Verified
Journal Article
Closed access
YCYanchao Chen
Yanchao Chen
WLWensong Liu
Wensong Liu
YFYueyuan Fan
Yueyuan Fan
DZDiming Zhang
Diming Zhang
LXLining Xing
Lining Xing
YLYuan-jiang Li
Yuan-jiang Li
Standard data-driven fault diagnosis methods typically rely on the assumption of identical label spaces across source and target domains. However, the emergence of novel failure modes in the target domain invalidates this premise, presenting a complex open-set fault diagnosis (OSFD) challenge where traditional closed-set techniques fail to generalize. To bridge this gap, we propose a novel Adaptive Prototype Network. This framework operates through a progressive two-stage paradigm: a warm-up phase that leverages prototype learning and domain adversarial training to establish coarse-grained alignment, followed by a refinement phase that optimizes the model via adaptive clustering and dual-risk minimization. Specifically, we introduce a dynamic sample weighting mechanism guided by a learnable unknown detector to suppress negative transfer. Extensive experiments on two distinct datasets demonstrate that proposed method effectively aligns shared classes while robustly segregating unknown faults, outperforming state-of-the-art baselines.