Zhejiang University of Technology

Pesticide residues in food matrices pose significant threats to global food safety. Conventional chromatographic methods, though highly reliable, are constrained by high costs, lengthy analysis times, and limited field deployability, restricting their use for rapid on-site monitoring across food supply chains. Nanomaterial-based fluorescent sensing platforms have emerged as promising analytical alternatives, leveraging the unique optical properties of quantum dots (QDs), carbon dots (CDs), Graphene quantum dots (GQDs), metal-organic frameworks (MOFs), and metal nanoclusters (MNCs). This review systematically examines recent advances in these platforms for pesticide detection, with particular emphasis on analytical chemistry perspectives spanning recognition mechanism design, signal transduction, and practical performance in complex food matrices. The structure-property relationships of major pesticide classes and their implications for molecular recognition are first discussed, followed by a systematic analysis of fluorescence modulation mechanisms, including photoinduced electron transfer (PET), Förster resonance energy transfer (FRET), inner filter effect (IFE), and aggregation-induced emission (AIE). Nanomaterial design strategies are further evaluated with respect to synthesis reproducibility, surface functionalization, and batch-to-batch consistency. The integration of artificial intelligence (AI) algorithms for automated signal processing and multi-analyte discrimination is highlighted as a promising approach for improving analytical robustness and quantitative accuracy. Recent advances in device engineering-including smartphone-based sensing platforms, paper-based microfluidic systems, and lateral flow immunoassay (LFIA)-are reviewed with respect to their potential for portable and field-deployable pesticide monitoring in complex food matrices. Finally, key challenges related to matrix interference, large-scale material production, and regulatory validation are discussed, with future directions emphasizing green nanomaterial synthesis, intelligent sensing architectures, and scalable analytical platforms to support distributed food safety monitoring networks.

Knowledge distillation (KD) can be broadly categorized into three paradigms: offline KD, online KD, and self-KD. However, these individual paradigms typically suffer from respective limitations in terms of adaptability, stability, and expressiveness. Meanwhile, their effective integration remains challenging due to heterogeneous supervision and the additional complexity introduced by multi-paradigm collaboration. To address these issues, we propose UKD, a Unified Knowledge Distillation framework that explicitly coordinates the three heterogeneous paradigms within a single optimization process for efficient knowledge transfer across diverse architectures and model capacities. Specifically, UKD introduces a lightweight proxy teacher to bridge the three paradigms and reduce capacity gap, which inherits the smoothed core knowledge from the pre-trained primary teacher, drives student-oriented knowledge adaptation and guides the internal consistency of student model. Moreover, UKD designs a dual-teacher collaboration strategy that dynamically adjusts teaching intensity of the powerful primary teacher and capacity-reduced proxy teacher based on the learning state of the student to improve the capacity adaptation of the model. To ensure efficiency, UKD further adopts backbone-sharing mechanisms with lightweight auxiliary modules to reduce training overhead. Extensive experiments on CIFAR-100 and ImageNet-1K across CNN, VIT, and MLP architectures demonstrate the effectiveness and efficiency of UKD, achieving gains of up to 11.38% and 2.68% with modest overhead, respectively. Code is available at https://anonymous.4open.science/r/UKD-ECD6.

Macrophage-based drug delivery systems (MDDSs) have garnered significant interest due to their inherent chemotactic properties. However, systematic methodologies for evaluating their migration and the effects of different drug-loading strategies remain insufficient. In this study, two structurally distinct MDDSs were developed to address this gap. Curcumin-loaded hollow silica nanoparticles (Cur-SHS) were utilized for intracellular endocytosis (Mφ-Cur-SHS), while their lipid-coated counterparts (Cur-SHS@lip) were anchored to the macrophage surface (Mφ-Cur-SHS@lip). A microfluidic chip platform mimicking normal and tumor-associated inflammatory microenvironments was established to dynamically evaluate the chemotactic behavior of both MDDSs. Assays demonstrated that the surface-bound system (Mφ-Cur-SHS@lip) achieved significantly higher targeting efficiency than the endocytosed system. The surface-anchored configuration reduced cytotoxicity and better preserved the intrinsic motility of macrophages, whereas excessive nanodrug loading impaired migration, underscoring the necessity of optimizing drug loading in macrophages. Furthermore, biodistribution study in tumor-bearing mice confirmed that both MDDSs effectively targeted tumor tissues. These findings provide critical insights into how structural configuration and drug loading influence the performance of MDDSs, and establish a robust microfluidic platform for the systematic evaluation and optimization of macrophage-based biomimetic delivery systems.