With the rapid evolution of Large Language Models (LLMs), generative recommendation is gradually reshaping the paradigm of recommender systems. However, most existing methods remain confined to the interaction-driven next-item prediction paradigm, struggling to keep pace with the latest evolving trends or address the diverse recommendation tasks along with business-specific requirements in real-world scenarios. To this end, we present SIGMA, a Semantic-Grounded Instruction-Driven Generative Multi-Task Recommender deployed at AliExpress. Specifically, we first ground item entities in a unified latent space capturing both general semantics and collaborative signals. Building upon this, we introduce a hybrid item tokenization method for both precise modeling and efficient generation. Moreover, we construct a large-scale multi-task supervised fine-tuning dataset empowering SIGMA to fulfill various recommendation demands via instruction-following. Finally, we design a three-step item generation procedure integrated with an adaptive probabilistic fusion mechanism to calibrate the output distributions based on task-specific requirements for recommendation accuracy and diversity. Extensive offline experiments and online A/B tests demonstrate the effectiveness of SIGMA across various real-world recommendation tasks.

To address the trade-off between security and efficiency in Industrial IoT, this article presents a memristive chaotic system and a feature-driven encryption architecture. First, a dynamic window decay (DWD) memristor incorporating a voltage-dependent window function is proposed. By coupling this memristor with the Ikeda map, a novel chaotic model termed the DWD-Ikeda map is developed. This model exhibits high-complexity chaotic behavior and serves as the pseudorandom source. Based on this, a 3-D plaintext-feature-driven encryption scheme is developed. Then, the architecture integrates fractional-wavelet-gradient feature extraction for dynamic parameter modulation and a feature-anchored chain reset diffusion mechanism to control error propagation. Finally, performance evaluations on standard benchmarks and the NEU-DET industrial dataset show a number of pixel change rate of 99.6% and a uniform average change intensity of 33.4%. For 256 × 256 images, the software encryption time is 0.20 s, while the FPGA hardware implementation achieves a throughput of 440 Mbps. These results verify the scheme’s suitability for secure real-time industrial image transmission.

ABSTRACT Aiming at the problem that the existing ultra‐short‐term wind power prediction methods lack consideration of the spatial correlation characteristics of wind farms, resulting in insufficient prediction accuracy, an ultra‐short‐term wind power prediction method based on spatiotemporal characteristics fusion is proposed in this article. First, the fluctuation difference of the time window of wind power is input into the K‐means clustering algorithm to cluster wind farms into several clusters based on power fluctuation similarity. Then, the principal component analysis algorithm is used to reduce the dimensionality of numerical weather prediction data combinations in different regions to reduce the impact of redundant information on modeling accuracy. Finally, a convolutional long‐short‐term memory neural network is designed to extract spatiotemporal features of wind power data and output prediction results. The experimental verification on 18 wind farms in a province in China shows that the proposed wind power prediction method has an average root mean square error of only 0.1257 and has certain applicability.

Power load forecasting is influenced by various factors, including meteorological, economic, and social factors. Considering all influencing factors will significantly increase the model complexity, affect accuracy and prediction timeliness. In addition, as time, region, and season change, the different influence factors will also change, which could lead a great impact on accuracy to existing prediction models. In order to improve the accuracy of real-time predictive analysis, this paper proposes a Dynamic Multi Indicator Fusion (DMIF) model for processing the calculation of real-time load impact indicators and adaptively predicting and adjusting the weights of the most influential indicators to reduce complex calculations and improve real-time prediction accuracy. In the final experiment, our model showed high prediction accuracy and fast calculation speed, while reducing information redundancy in multiple indicators. Therefore, in the context of smart grids, this method has practical application value for the stable operation of power grid systems.

Summary With the rapid expansion of global offshore wind power market, the research on improving the full life cycle income and reducing the construction and operation and maintenance costs has attracted the attention of scholars in the industry. In view of the different aging degree and maintenance cycle of wind turbines, this paper studies the optimized design of patrol path for offshore wind farms based on genetic algorithm (GA) and particle swarm optimization (PSO) with traveling salesman problem (TSP). Firstly, the problem of patrol routing planning in offshore wind farms is described as the traveling salesman problem of shortest route optimization. Secondly, the GA and PSO algorithms are simulated and verified separately, and the patrol path distance is taken as the objective function. Finally, through simulation experiments, the optimized patrol path performances of PSO and GA are compared, which can help to find a shortest route and reduce the operation and maintenance costs.

Summary The experienced doctors can alleviate symptoms such as headaches, insomnia, anxiety, and depression by observing the patient's ears and massaging specific areas. In order to achieve remote ear condition diagnosis and guide patients to massage their ears independently through the network, patients can use their mobile phones to take and send photos of ears to doctors. However, due to significant differences in the clarity of photos taken by different mobile phones, as well as susceptibility to haze, lighting, jitter, and low pixels, the quality of photos is poor, which affects the accuracy of remote diagnosis by doctors. This study adopted an image preprocessing method based on He Kaiming's dark channel prior dehazing method to enhance the original ear images captured by mobile phones. The dehazing algorithm was used to remove the haze effect of the ear images, improving image quality and contrast, making the wrinkles, protrusions, pigmentation and other areas of the ear more obvious. The experiment has showed the comparison by adjusting weight from 15% to 95% between two methods—dark channel prior method and the dark channel prior method after preprocessing, which has proven the effectiveness of dehazing method in human ear images taken by mobile phones. The image quality after preprocessing and dehazing is widely recognized and accepted by doctors at hospitals in Hangzhou, China.

Offshore wind power is the fastest growing clean energy in the world. Since the harsh operating environment, the operation and maintenance costs are high, therefore, the digital intelligent solutions are very necessary. This article proposed a tide forecast method based on random forests. The random forests model was trained using the tide level data from a tide gauge station for 3 years, and the actual monitoring data was verified by experimental simulation. The results proved that the proposed method can effectively forecast the non-astronomical tide, which was better than harmonic analysis method. The proposed method can provide a more reliable tidal information for the operation and maintenance task planning in offshore wind farms, so as to reduce the unnecessary number of operation and maintenance plans, and reduce the cost.

The morphological similarity of anatomical structures is essential to the study of the species evolution. In this paper, we investigate the unsupervised shape similarity analysis by a random-forest-based metric. The dense continuous deformation fields are employed as the shape descriptors. The forest is built when given the unlabeled deformation fields, where the leaves can be seen as an optimal clustering of the data set. The salient region is defined based on the dominant feature channels determined by the forests. The pairwise shape distance is computed efficiently with just binary comparisons stored in tree branches. We have applied our method to several skeletal data sets, including the skulls, teeth, radii, and metatarsals. Our experiments demonstrate the proposed method can handle the taxonomic classification effectively.