Despite the encouraging achievements, the practical application of recommendation systems still faces two key issues. The first is how to better understand the multimodal real-time requests that are the more mainstream request behavior in industrial scenarios; the other is how to effectively capture users' dynamic needs that change with temporal and spatial conditions. The breakthroughs in text understanding and generation capabilities of Large Language Models (LLMs) have demonstrated their tremendous potential in precise recommendation systems, particularly through the enhancement of the understanding of user intent. To address these issues, we propose a novel Spatial-Temporal Multimodal LLM for generative recommendation. Specifically, on the basis of the behavior data constructed from Alipay, spatial-temporal knowledge-guided fine-tuning module is proposed to capture specific needs in user real-time requests. Furthermore, a preference discovery module is developed to learn user preferences in visual queries from multimodal request perspective. Meanwhile, a personalized recommendation module is designed to aggregate spatial-temporal knowledge and user preferences for generative recommendation. Experimental results on a real-world deployed generative recommendation task from the ‘Explore' scenario in Alipay have demonstrated the effectiveness of the proposed framework.

Summary When it comes to the dehazing method based on the ambient light scattering model, there are ill‐conditioning problems of great concern. Therefore, the inaccurate estimation of the global atmospheric light and transmittance can bring about hue error phenomena, halo artifacts, and image noises. In this paper, we present an effective and efficient approach to solve above problems. Firstly, we adopt single‐threshold segmentation method and quad‐tree algorithms, which can break through the limitation that the dark channel prior theory (DCP) is not suitable for sky region, to quickly and accurately locate and estimate the global atmospheric light. Secondly, we design a new deep convolutional network architecture, which can effectively suppress the halo effect of image edges and restrain image noises, to optimize transmittance. Numerous of outdoor and synthesis haze image test results indicate that the optimized approach could shorten the processing time greatly. Besides, with more accurate global atmospheric light value and transmittance, the image noise and the halo artifact can be effectively suppressed and removed respectively, the hue error phenomenon improved, and the image contrast significantly enhanced in the dehazing results.