Social bot detection is pivotal for safeguarding the integrity of online information ecosystems. Although recent graph neural network (GNN) solutions achieve strong results, they remain hindered by two practical challenges: (i) severe class imbalance arising from the high cost of generating bots, and (ii) topological noise introduced by bots that skillfully mimic human behavior and forge deceptive links. We propose the Reinforcement-guided graph Augmentation social Bot detector (RABot), a multi-granularity graph-augmentation framework that addresses both issues in a unified manner. RABot employs a neighborhood-aware oversampling strategy that linearly interpolates minority-class embeddings within local subgraphs, thereby stabilizing the decision boundary under low-resource regimes. Concurrently, a reinforcement-learning-driven edge-filtering module combines similarity-based edge features with adaptive threshold optimization to excise spurious interactions during message passing, yielding a cleaner topology. Extensive experiments on three real-world benchmarks and four GNN backbones demonstrate that RABot consistently surpasses state-of-the-art baselines. In addition, since its augmentation and filtering modules are orthogonal to the underlying architecture, RABot can be seamlessly integrated into existing GNN pipelines to boost performance with minimal overhead.

The widespread and often excessive use of pesticides presents serious risks to human health and environmental safety, calling for IoT-enabled monitoring in real agricultural environments. Current detection methods face challenges in handling diverse pesticide compounds, operating portably, and extracting discriminative signal features. To overcome these limitations, we presentHornbill+, a portable and high-precision electrochemical sensing system. By combining NFC technology with electrochemical biosensing,Hornbill+supports accurate, contactless, and multi-pesticide classification in field-friendly settings. The principle ofHornbill+involves recording electron transfer behaviors of selected biological materials under varying electrode potentials, producing time-variant electrochemical fingerprints that reflect distinct reaction signatures for different pesticides. To implement this approach, we developed a dual-channel fully differential potentiostat integrated into a low-power NFC tag, using DPV as the electrochemical readout method to enhance detection sensitivity. To enhance accuracy in complex real-world scenarios, we integrated a pyramid attention mechanism into a deep learning model for interpreting electrochemical dynamics.Hornbill+achieves over 93% average accuracy across 18 pesticides, five concentrations, and nine mixtures, surpassing existing techniques in both precision and coverage.

Automatic modulation recognition plays a critical role in military applications, particularly in electronic warfare, spectrum surveillance, and secure communication systems. The precise identification of signal modulation modes is crucial for ensuring the efficacy, security, and efficiency of communications. Given the problems of limited feature extraction capability and performance degradation when dealing with low signal-to-noise ratio signals, this work proposes a novel model architecture that combines the Transformer with relative position encoding and separable convolution in the complex domain. The network can directly process the complex representation of signals to capture features in the time-frequency domain and enhance the ability to recognize complex signals. This method introduces relative position encoding in the complex domain into the Transformer framework, which uses complex attention mechanisms and adaptive position encoding to enhance the model’s long-range modeling capability. At the same time, it improves computational efficiency and local feature extraction capability by introducing separable convolution layers. Then, we construct an attention-driven feature fusion module, which can automatically adjust the weight ratio between features to achieve the optimal combination of features. The model shows excellent classification performance under various signal-to-noise ratio conditions in RadioML2016.10a and RadioML2016.10b datasets, especially in low signal-to-noise ratio environments, which is significantly improved compared to other methods. The research not only provides a new solution for AMR tasks but also expands new ideas for applying the Transformer model in signal processing.

The accurate acquisition of land surface temperature (LST) on complex mountainous surfaces has always been a difficult problem and hot topic in thermal infrared remote sensing inversion, and the uncertainty caused by the radiation angle effect is one of the important factors hindering the accurate inversion of LST. Researchers have proposed a variety of models to simulate and eliminate the influence of the angle effect, among which, the kernel-driven model has a bright prospect of development. However, fewer studies have been conducted to observe the effects of terrain and land cover on thermal radiation directionality (TRD) properties based on measured data. This paper intends to carry out observations on a small spatial scale of a complex mountainous area using an unmanned aerial vehicle (UAV) to investigate the specific effects of different slopes, aspects, and land cover on the TRD characteristics. The measured results show that the intensity of thermal radiation anisotropy is actively correlated with the complexity of surface structure, and the influence of slope on TRD bias is in the form of a staged “S”, where its intensity increases slowly and then rapidly before slowing down again; the dispersion of thermal radiation in TRDs of different aspects is affected by the duration and intensity of solar radiation, and there is a time lag effect. Meanwhile, in order to evaluate the accuracy of different radiation directionality models, this paper further evaluates the currently more recognized kernel-drive model named LSF-Chen and the thermal equivalent slope kernel-driven (TESKD) model based on the TRD measurements from UAVs. The results show that the correlation coefficients of simulation results and measurements are all greater than 0.6, and the RMSEs are all less than 2K, and that the two methods both have a better simulation of the TRD effect, but the TESKD is better overall in terms of accuracy and methodological details. Through this study, a new method of applying UAVs to capture the thermal direction of the complex surface in mountainous areas is proposed, which provides methodological support for the extraction and accurate simulation of the TRD characteristics on the complex mountainous areas.

Land surface emissivity (LSE) is a key parameter for estimating longwave radiation of land surface, and mounts of the satellite-based LSE products have been released, generally coupled with Land surface temperature (LST) products. However, few research focus on validation of remote sensing LSE products, particularly over complex and heterogeneous mountainous surface. In this study, two-year field experiments designed for the LSE observation was implemented over typical mountainous regions of southwestern China, using a Model 102 hand-portable Fourier-transform Infrared (FTIR) spectrometer. Through an optimized sampling method, mixed pixel emissivity measurements were obtained to systematically evaluate the widely used Moderate Resolution Imaging Spectroradiometer (MODIS) Collection 6.1 level-3 daily LSE products, including MOD/MYD11A1 and MOD/MYD21A1 and the Landsat8/9 Collection 2 level 2 LSE products. In this study, the Landsat LSE product shows slightly higher accuracy, with the mean Absolute Bias (Abs_Bias) of 0.0104, compared to MxD11A1 (Abs_Bias = 0.0118) and MxD21A1 (Abs_Bias = 0.0125). Between the two MODIS products, MxD11A1 tends to overestimate LSE with the mean Bias of 0.0115, while MxD21A1 shows a slight underestimation, with the mean Bias of –0.0022. Regarding sensor differences, MxD11A1 shows negligible discrepancies between Terra and Aqua platforms (Bias, Abs_Bias, RMSE < 0.0001) whereas MYD21A1 exhibits larger errors than MOD21A1, with the Abs_Bias and RMSE higher by 0.0029 and 0.0043, respectively, indicating that Terra products generally perform better than Aqua. For daytime and nighttime comparisons, MOD21A1 exhibits minor differences, with Abs_Bias values of 0.0112 and 0.0110, whereas MYD21A1 nighttime product performs better than daytime counterpart, with lower Abs_Bias (0.0122 vs. 0.0157) and RMSE (0.0149 vs. 0.0209). While the Landsat product achieves slightly better overall absolute accuracy, it exhibits spatial artifacts that result in underestimation in affected regions and slight overestimation in unaffected areas. These artifacts also limit its sensitivity to temporal variation. Overall, this study provides a reliable accuracy reference for MODIS and Landsat LSE products over complex and heterogeneous mountainous surfaces, supporting their application and the future improvement of product quality.

Pesticide overuse poses significant risks to human health and environmental integrity. Addressing the limitations of existing approaches, which struggle with the diversity of pesticide compounds, portability issues, and environmental sensitivity, this paper introduces Hornbill. A wireless and battery-free electrochemical bio-tag that integrates the advantages of NFC technology with electrochemical biosensors for portable, precise, and touchless multi-pesticide detection. The basic idea of Hornbill is comparing the distinct electrochemical responses between a pair of biological receptors and different pesticides to construct a unique set of feature fingerprints to make multi-pesticide sensing feasible. To incorporate this idea within small NFC tags, we reengineer the electrochemical sensor, spanning the antenna to the voltage regulator. Additionally, to improve the system's sensitivity and environmental robustness, we carefully design the electrodes by combining microelectrode technology and materials science. Experiments with 9 different pesticides show that Hornbill achieves a mean accuracy of 93% in different concentration environments and its sensitivity and robustness surpass that of commercial electrochemical sensors.

With the increasing demand for low power and storage consumption in mobile platforms, wearable devices, and Internet of Things devices, how to better apply lightweight neural networks in many edge computing scenarios and resource-limited settings is still facing challenges. This paper first proposes a novel binary convolution structure based on the time-domain to reduce resource and power consumption for the convolution process. Furthermore, through the joint design of convolution, batch normalization, and activation function in the time-domain, we propose a full-BNN model and hardware architecture, which keeps the values of all intermediate results as one bit to reduce storage requirements by 75%. Then, we optimize the above design with spatial and temporal parallelism to improve the overall computing efficiency. Finally, we built an accelerating system and take the MNIST data set as an example to test the optimized architecture on the DSP + FPGA platform. The results show that the model can be used as a neural network acceleration unit with low storage requirement and low power consumption for classification tasks with a small loss of accuracy. The joint design method in the time-domain may further inspire other computing architectures.