The rapid proliferation of the Internet of Things (IoT) has created an urgent need for on-device intelligence that balances high computational demands with stringent energy constraints. Existing edge inference frameworks struggle to deploy generative artificial intelligence (AI) models efficiently on low-power devices, often sacrificing fidelity for efficiency or lacking adaptability to dynamic conditions. To address this gap, we propose a generative AI edge inference framework integrating lightweight architecture compression, adaptive quantization, and energy-aware scheduling. Extensive experiments on CIFAR-10, Tiny-ImageNet, and IoT-SensorStream show that our method reduces energy consumption by up to 31% and inference latency by 27% compared with state-of-the-art baselines, while consistently improving generative quality. Robustness tests further confirm resilience under noise, cross-task, and cross-dataset conditions, and ablation studies validate the necessity of each module. Finally, deployment in a hospital IoT laboratory demonstrates real-world feasibility. These results highlight both the theoretical contribution of unifying compression, quantization, and scheduling, and the practical potential for sustainable, scalable, and reliable deployment of generative AI in diverse IoT ecosystems.

Cognitive diagnosis is an important component of adaptive learning, as it infers learners’ latent knowledge states and enables tailored feedback. However, existing approaches often emphasize sequential modeling or latent factorization, while insufficiently incorporating curriculum structures that embody prerequisite relations. This gap constrains both predictive accuracy and pedagogical interpretability. To address this limitation, we propose a Curriculum-Aware Graph Neural Cognitive Diagnosis (CA-GNCD) framework that integrates curriculum priors into graph-based neural modeling. The framework combines graph representation learning, knowledge-prior fusion, and interpretability constraints to jointly capture relational dependencies among concepts and individual learner trajectories. Experiments on three widely used benchmark datasets, ASSISTments2017, EdNet-KT1, and Eedi, show that CA-GNCD achieves consistent improvements over classical probabilistic, psychometric, and recent neural baselines. On average, it improves AUC by more than 4.5 percentage points and exhibits relatively faster convergence, greater robustness to noisy conditions, and stronger cross-domain generalization. These results suggest that aligning diagnostic predictions with curriculum structures can enhance interpretability and reliability, offering implications for personalized learning support. While promising, further validation in diverse educational contexts is required to establish the generalizability and practical deployment of the proposed framework.

To address the imbalance in energy supply and demand across different regions and seasons, the thermochemical conversion process was selected to efficiently utilize surplus energy. In the search for suitable novel materials, this study developed a porous matrix “in-salt” composite using a carbonized metal-organic framework as the carrier and LiCl as the primary reactant. When exposed to water vapor, the innovative material enabled both adsorption and desorption of water vapor, leading to the release and storage of thermal energy, thereby achieving effective energy storage. Using Zn(NO₃)₂·6H₂O and Co(NO₃)₂·6H₂O as metal ion sources and 2-methylimidazole as the ligand, bimetallic zeolitic imidazolate frameworks (BMZIFs) were fabricated via the liquid-phase precipitation method. The composite specimen prepared at a carbonization temperature of 1000 °C with a 20% LiCl mass concentration exhibited the most promising thermal storage performance, achieving the highest capacity, with a final water loss of 53.56% and a stable water adsorption capacity of about 0.831 g·g⁻¹.

Perovskite solar cells (PSCs) rarely report, on a single-device platform, concurrent gains in optoelectronic efficiency and buried-interface mechanical robustness—two prerequisites for flexible and roll-to-roll (R2R) integration. We engineered a nanoplate-structured fluorine-doped tin oxide (NP-FTO) front electrode that couples light management with three-dimensional interfacial anchoring, and we quantified both photovoltaic (PV) and nanomechanical metrics on the same device stack. Relative to planar FTO, the NP-FTO PSCs achieved PCE of up to 25.65%, with simultaneous improvements in Voc (to 1.196 V), Jsc (up to 26.35 mA cm⁻²), and FF (to 82.65%). Nanoindentation revealed a ~28% increase in reduced modulus and >70% higher hardness, accompanied by a ~32% reduction in maximum indentation depth, indicating enhanced load-bearing capacity consistent with the observed FF gains. The low-temperature, solution-compatible NP-FTO interface is amenable to R2R manufacturing and flexible substrates, offering a unified route to bridge high PCE with reinforced interfacial mechanics toward integration-ready perovskite modules.

With the development of decentralized artificial intelligence (AI) networks, federated learning (FL) has received extensive attention for its ability to enable collaborative modeling without sharing raw data. However, existing methods are prone to convergence instability under non-independent and identically distributed (non-IID) conditions, lack robustness in adversarial settings, and have not yet sufficiently addressed fairness and incentive issues in multi-source heterogeneous environments. This paper proposes a Symmetry-Preserving Federated Learning (SPFL) framework that integrates blockchain auditing and fairness-aware incentive mechanisms. At the optimization layer, the framework employs group-theoretic regularization to maintain parameter symmetry and mitigate gradient conflicts; at the system layer, it leverages blockchain ledgers and smart contracts to verify and trace client updates; and at the incentive layer, it allocates rewards based on approximate Shapley values to ensure that the contributions of weaker clients are recognized. Experiments conducted on four datasets, MIMIC-IV ECG, AG News-Large, FEMNIST + Sketch, and IoT-SensorStream, show that SPFL improves average accuracy by about 7.7% compared to FedAvg, increases Jain’s Fairness Index by 0.05–0.06 compared to FairFed, and still maintains around 80% performance in the presence of 30% Byzantine clients. Convergence experiments further demonstrate that SPFL reduces the number of required rounds by about 30% compared to FedProx and exhibits lower performance degradation under high-noise conditions. These results confirm SPFL’s improvements in fairness and robustness, highlighting its application value in multi-source heterogeneous scenarios such as medical diagnosis, financial risk management, and IoT sensing.

Transparent and comparable evaluation of ecological restoration outcomes is essential for advancing performance-based environmental governance and ESG-aligned ecological compensation. However, existing grassland monitoring approaches in semi-arid regions often rely on single vegetation indices, which fail to capture ecosystem structure, functional recovery, and temporal dynamics. To address these limitations, this study proposes a process-oriented Restoration Index (RI) based on multi-source remote sensing data. By integrating spectral, textural, and phenological indicators, together with topographic and climatic factors, derived from Sentinel-2 and Landsat time-series imagery, the framework characterizes vegetation productivity, community structure, and seasonal ecological processes within a unified analytical framework. A case study in the Xilingol grassland of Inner Mongolia shows that different management strategies, including grazing exclusion, reseeding, and rotational grazing, are associated with distinct restoration trajectories and recovery performance. The results indicate that the RI captures both spatial heterogeneity and temporal evolution of ecosystem recovery, while the normalization procedure improves the relative comparability of restoration assessment results within the adopted framework. Quantitative evaluation shows positive agreement with field observations, providing preliminary support for the applicability of the approach within the study area. Overall, the RI framework provides a scalable and policy-relevant basis for ecological restoration assessment and may support ecological compensation evaluation, environmental auditing, and more transparent restoration governance.