While modern recommender systems are instrumental in navigating information abundance, they remain fundamentally limited by static user modeling and reactive decision-making paradigms. Current large language model (LLM)-based agents inherit these shortcomings through their overreliance on heuristic pattern matching, yielding recommendations prone to shallow correlation bias, limited causal inference, and brittleness in sparse-data scenarios. We introduce STARec, a slow-thinking augmented agent framework that endows recommender systems with autonomous deliberative reasoning capabilities. Each user is modeled as an agent with parallel cognitions: fast response for immediate interactions and slow reasoning that performs chain-of-thought rationales. To cultivate intrinsic slow thinking, we develop anchored reinforcement training-a two-stage paradigm combining structured knowledge distillation from advanced reasoning models with preference-aligned reward shaping. This hybrid approach scaffolds agents in acquiring foundational capabilities (preference summarization, rationale generation) while enabling dynamic policy adaptation through simulated feedback loops. Experiments on MovieLens 1M and Amazon CDs benchmarks demonstrate that STARec achieves substantial performance gains compared with state-of-the-art baselines, despite using only 0.4% of the full training data.

Modern online advertising systems often involve a substantial number of advertisers in each auction, which results in scalability issues. To address this challenge, two-stage auctions have been designed and implemented in practice. These auctions enable efficient allocation of ad slots among numerous candidate advertisers in a short response time. This approach employs a fast yet coarse model in the first stage to select a small subset of advertisers, followed by a slow, more refined model to determine the final winners. However, existing two-stage auction mechanisms primarily focus on optimizing welfare, overlooking other critical objectives of the platform, such as revenue.

Monitoring Arctic sea ice drift is essential for understanding climate change. Currently, large-scale observed sea ice drift datasets primarily rely on single-sensor remote sensing data, which suffer from low spatial resolution or poor spatial continuity. To address these limitations, this study proposes a sea ice drift retrieval framework based on multi-sensor data, utilizing the complementary sea ice drift information derived from passive microwave radiometers and medium-resolution optical sensors. The proposed framework employs the maximum cross-correlation (MCC) based pattern-matching method to obtain sea ice drift fields from coarse-resolution Fengyun-3D (FY-3D) Microwave Radiation Imager (MWRI) data, and an A-KAZE-based feature-tracking method to extract sea ice motion vectors from FY-3D Medium-Resolution Spectral Imager II (MERSI-II) data. Finally, the sea ice drift vectors from different sensors are merged using the Co-Kriging algorithm to obtain the final sea ice drift result. The effectiveness of the proposed framework was assessed by comparing displacements from 166 buoys with the retrieved vectors derived from FY-3D single-sensor and multi-sensor data, as well as an existing sea ice drift product (Ifremer-AMSR2) collected in the Beaufort Sea, the East Siberian Sea, and the Fram Strait. The results demonstrate the proposed framework’s ability to retrieve fine-resolution (i.e., 1 km) and spatially continuous sea ice drift fields in areas where vectors from fine-resolution data can be obtained. The overall mean absolute errors (MAEs) of the merged sea ice motion vectors are 0.76 km/day for speed and 4.53° for angle, exhibiting superior drift accuracy to Ifremer-AMSR2 in areas covered by MERSI-II vectors.

In recent years, deep-learning-based hyperspectral image (HSI) classification methods have achieved significant development and gradually become widely applied. The existing advanced methods can achieve near-saturation performance with sufficient labels in a closed-set environment (CSE), i.e., training set and test set are all known categories of ground objects. However, the real world is usually open because of the diversity of land covers, i.e., test-set exists unknown categories that are not labeled in the training set. Therefore, the prevalent advanced CSE methods still cannot effectively and robustly handle unknown categories of ground objects in an open-set environment (OSE). Therefore, we propose a spectral-spatial MLP-like network with reciprocal points learning (SSMLP-RPL) to improve the performance of open-set HSI classification. First, a feature learning framework based on reciprocal points learning (RPL) is constructed to model the extra-category space and reduce the risk of open space. The learned feature space enables to enlarge the distance between the known and unknown categories. Besides, we further propose to utilize a learnable dynamic threshold of each known category to effectively distinguish the unknown categories and improve open performance of the model. Second, to enhance the capacity of feature learning, a spectral-spatial MLP-like network (SSMLP) is designed to capture the spectral-spatial feature merely with a series of fully-connected (FC) layers, which mainly involve SpeFC and SpaFC two modules. Among them, the SpaFC module enables to model spacial semantics, and the SpeFC module enables to model long-distance spectral dependence. Extensive experiments on three benchmark HSIs show that SSMLP-RPL has a competitive performance both in CSE and OSE and even surpasses currently advanced closed-set and open-set HSI classification methods. As an end-to-end HSI classification framework of MLP-backbone, SSMLP network can compete with the advanced works based on CNN and transformer. The code will be open at: https://github.com/sssssyf/SSMLP-RPL.

Time is an important index of the search and rescue (SAR) ability. In this paper, a method for forecasting the location and probability of patrol vessel which is responsible for SAR in South China Sea is presented. In this study, based on the density of normal sailing ships passing the area and the distance from ports of each base port, the distribution probability of patrol vessels in South China Sea were obtained. The SAR radius and times before and after bases setting in Xisha and Nansha were calculated based on the proposed method. The results indicated that the SAR situation in South China Sea changed a lot before and after bases setting in Xisha and Nansha.

Recently, the field of hyperspectral image (HSI) classification is dominated by deep learning-based methods. However, training deep learning models usually needs a large number of labeled samples to optimize thousands of parameters. In this article, a deep multiview learning method is proposed to deal with the small sample problem of HSI. First, two views of an HSI scene are constructed by applying principal component analysis to different bands. Second, a deep residual network is designed to embed the different views of a sample to a latent space. The designed deep residual network is trained by maximizing agreement between differently augmented views of the same data sample via a contrastive loss in the latent space. Note that the training procedure of the designed deep residual network does not use labeled information. Therefore, the proposed method belongs to the category of unsupervised learning, which could alleviate the lack of labeled training samples. Finally, a conventional machine learning method (e.g., support vector machine) is used to complete the classification task in the learned latent space. To demonstrate the effectiveness of the proposed method, extensive experiments are carried on four widely used hyperspectral data sets. The experimental results demonstrate that the proposed method could improve the classification accuracy with small samples.

The nighttime light (NTL) on the surface of earth is a great symbol indicating the changes made by human activities such as urban lighting, gas flaring and other living or producing activities. Therefore NTL remote sensing can be used for discovering and monitoring the human activities connecting NTLs. In this paper, NTL remote sensing has been successfully used in agricultural area, based on the light supply system of dragon fruit plantations. A method of extracting dragon fruit plantations had been proposed, and exemplified in Binh Thuan province of Vietnam.

Deep learning methods have recently been successfully explored for hyperspectral image (HSI) classification. However, training a deep-learning classifier notoriously requires hundreds or thousands of labeled samples. In this paper, a deep few-shot learning method is proposed to address the small sample size problem of HSI classification. There are three novel strategies in the proposed algorithm. First, spectral–spatial features are extracted to reduce the labeling uncertainty via a deep residual 3-D convolutional neural network. Second, the network is trained by episodes to learn a metric space where samples from the same class are close and those from different classes are far. Finally, the testing samples are classified by a nearest neighbor classifier in the learned metric space. The key idea is that the designed network learns a metric space from the training data set. Furthermore, such metric space could generalize to the classes of the testing data set. Note that the classes of the testing data set are not seen in the training data set. Four widely used HSI data sets were used to assess the performance of the proposed algorithm. The experimental results indicate that the proposed method can achieve better classification accuracy than the conventional semisupervised methods with only a few labeled samples.

Waterway is a kind of narrow water area which connects mainland or islands. It acts the role of channel or passage, as the only path through some area. Usually, waterways exist with the islands nearby in the sea. So the spatial relationship between islands and waterways is very important. In this paper, spatial interaction model is used to describe the relationship and the importance of the location of islands in research area is analyzed.

A visual circle based image registration algorithm for optical and SAR imagery is proposed in this paper. The visual circles with robustness to angle differences and flexible resolution build a bridge between the reference and sensed images with different resolutions and angles, and combine the spatial features and correlation measures together for the feature matching. Two universal correlation measures comprising normalized cross correlation coefficient (NCCC) and normalized mutual information (NMI) are adopted respectively for registration between Radarsat-2 and ASTER images. In order to testify the visual circle's robustness to rotation, the visual squares and circles were constructed meanwhile and used to register the rotated sensed images with different angles respectively. The results proved that the total RMS error of the proposed algorithm was less than one pixel and the circle had higher robustness to angle differences and better performance than that of the square.

When applying SIFT algorithm in registration between multi-optical imagery, few matching couples are obtained. Aiming to this problem, a new algorithm named normalized SIFT based on visual matching window and structural information is proposed in this paper. By constructing the visual matching window, the scale similarity and matching probability between SIFT features are significantly increased. The structure consistency-based matching couples' inspection method is carried out to remove incorrect matching points, which increases the registration accuracy. Some multi-sensor remote sensing images with big differences in acquired angle and resolution were tested. The results indicate that many matching couples which distribute equally were extracted and the precision is lower than 1 pixel.