Temporal Action Detection (TAD) aims to identify action boundaries and their corresponding categories in untrimmed videos, playing a crucial role in long-video understanding. Prior works often struggle to balance the trade-off between capturing long-range dependencies and ensuring computational efficiency. Recently, the state space model Mamba has exhibited impressive capabilities and efficiency in long-term sequence modeling. However, current methods based on Mamba generally lack a unified framework to simultaneously address the redundancy of long-duration actions and the boundary sensitivity of short-duration actions—limitations that largely stem from Mamba’s reliance on limited state representations and its unidirectional modeling. To tackle the aforementioned challenges, we propose DilatedTAD, a novel TAD framework with an expanded receptive field. DilatedTAD leverages the Inter-Parallel DIM component (InterDIM) to integrate multi-scale temporal information, enabling a better trade-off between short-duration and long-duration action detection. InterDIM is built upon our proposed Dilated Mamba (DIM), where multiple DIM branches with different dilation rates are designed to focus on actions of varying durations. Specifically, DIM introduces a novel use of dilation to skip redundant temporal information, thereby enhancing the model’s focus on crucial boundary features. Additionally, a bidirectional modeling design is adopted in DIM to compensate for the lack of future temporal context in the original Mamba architecture. Extensive experiments show that DilatedTAD outperforms state-of-the-art methods on multiple datasets, achieving mAPs of 74.9% (THUMOS14), 42.90% (ActivityNet 1.3), 45.0% (HACS), and 26.3% and 24.3% (EPIC-Kitchens 100). Our code will be publicly available.

This paper proposes an innovative optimization framework, Dynamic Graph Neural Evolution (DGNE), integrating Graph Neural Networks (GNNs) with Evolutionary Algorithms (EAs). Building on the foundation of Graph Neural Evolution (GNE), DGNE introduces a dynamic filtering mechanism and adaptive Gaussian sampling functions to dynamically adjust the population distribution during the optimization process, achieving a balance between global exploration and local exploitation. By emphasizing high-frequency information in the early stages to enhance population diversity and low-frequency information in the later stages to promote convergence, DGNE effectively optimizes the search process. Experiments were conducted on the CEC2017 benchmark suite across 30, 50, and 100 dimensions, comparing DGNE with advanced algorithms (LSHADE, LSHADE_cnEpSin, MadDE, SaDE, EA4eig) and classic algorithms (DE and CMA-ES). Statistical analyses using the Wilcoxon rank-sum test and Friedman mean rank test demonstrate that DGNE achieves the best average ranking in 50 and 100 dimensions and ranks third in 30 dimensions. However, it achieves the highest overall average performance ranking across all dimensions, showcasing its stability and significant advantages in different scenarios. While its performance in low-dimensional tasks is slightly less competitive compared to high-dimensional ones, DGNE still exhibits strong competitiveness. Additionally, we explored the impact of population size. DGNE was evaluated across population sizes of 20, 30, 50, and 100. The results highlight DGNE’s robustness, maintaining competitive rankings across all population sizes, with top rankings for smaller population sizes (20 and 30) and strong results at larger sizes (50 and 100). These findings confirm DGNE’s adaptability to varying configurations and further validate its effectiveness as a robust optimization framework. Overall, DGNE demonstrates great potential as an optimization method, offering a promising direction for further research and applications in artificial intelligence and optimization fields. Its ability to remain competitive across diverse tasks and configurations underscores its versatility and scalability.

Semi-supervised semantic segmentation has attracted considerable attention for its ability to mitigate the reliance on extensive labeled data. However, existing consistency regularization methods only utilize high certain pixels with prediction confidence surpassing a fixed threshold for training, failing to fully leverage the potential supervisory information within the network. Therefore, this paper proposes the Uncertainty-participation Context Consistency Learning (UCCL) method to explore richer supervisory signals. Specifically, we first design the semantic backpropagation update (SBU) strategy to fully exploit the knowledge from uncertain pixel regions, enabling the model to learn consistent pixel-level semantic information from those areas. Furthermore, we propose the class-aware knowledge regulation (CKR) module to facilitate the regulation of class-level semantic features across different augmented views, promoting consistent learning of class-level semantic information within the encoder. Experimental results on two public benchmarks demonstrate that our proposed method achieves state-of-the-art performance. Our code is available at https://github.com/YUKEKEJAN/UCCL.

Recent researches have yielded promising results by integrating online action detection and action anticipation tasks to explore the correlations between past, present and future. However, these approaches treat incomplete historical information equally and neglect intrinsic connections between actions, resulting in a limited perception of the throughout evolution. To address this limitation, we reconsider the patterns and dependencies in event evolution, innovatively constructing a comprehensive deductive process that inscribes the entire temporal spectrum via procedural features. Here, we propose the Throughout Procedural Transformer (TPT) comprising Procedural History Evolution Encoder and Progressive Deduction Decoder, to thoroughly span the entirety of time from history to the future through procedural modeling. TPT utilizes long-term procedural history acquired through procedure sampling to model long-term procedural future, thereby enhancing cognitive inference ability by enriching short-term history and short-term future with a broad grasp of throughout event evolution. We conduct extensive experiments to evaluate TPT on five demanding benchmarks THUMOS’14, TVSeries, FineAction, HACS and EPIC-Kitchens-100 for online action detection and anticipation tasks, demonstrating significant improvements over existing methods.

Matching molecular analogues is a computational chemistry and bioinformatics research issue which is used to identify molecules that are structurally or functionally similar to a target molecule. Recent studies on matching analogous molecules have predominantly concentrated on enhancing effectiveness, often sidelining computational efficiency, particularly in contexts of low computational resources. This oversight poses challenges in many real applications (e.g., drug discovery, catalyst generation and so forth). To tackle this issue, we propose a general strategy named MapLE, aiming to promptly match analogous molecules with low computational resources by multi-metrics evaluation. Experimental evaluation conducted on a public biomolecular dataset validates the excellent and efficient performance of the proposed strategy.

The Internet of Things (IoT) is making the world around us smarter and more convenient. However, its extensive application has rendered security problems, such as the privacy of sensitive data, increasingly serious. Encryption provides an effective and important means of protecting data privacy in the IoT. Because of resource limitations, to achieve high efficiency IoT devices require an encryption scheme that ensures that the encryption phase does not incur a heavy data transmission overhead. By virtue of its many advantages, the use of public-key encryption in current applications is widespread. Identity-based public-key encryption (IBE) removes the obstacle raised by the sophisticated certificate management required by other schemes, and its efficiency renders it more suitable for application in the IoT. However, a problem that needs to be solved in IBE is the revocation of a user whose private key may have been exposed. In this article, we present an efficient and practical IBE scheme having a revocation functionality to preserve data privacy in IoT applications. Elements in the system, such as sensors and actuators, can exchange encrypted data directly or via a cloud server. If a private key is compromised, the private key generator can revoke its user. The security of our proposed scheme can be proved based on SM9 encryption and the bilinear Diffie–Hellman problem.

Object detection, as one of the most important environment perception tasks for traffic safety in intelligent transportation systems, has been widely investigated recently. However, most of the researches focus on the fully supervised scenario, and inevitably lead to model failure. With the continuous development of Zero-Shot Learning (ZSL) models, Generalized Zero-Shot Detection (GZSD) has attracted great attention due to its ability of detecting unseen objects. Many researchers tend to map the detected visual features to semantic attributes and then separate seen and unseen domains during inference. But they have ignore that the generative methods generally have higher performance than these visual-semantic mapping methods, and they have been confirmed from previous GZSL methods. In order to make up for the vacancy of GZSD in the generative methods, we propose an idea of using curriculum learning to generate more precise unseen visual features. And with the excellent performance of WGAN-based method in sample synthesis, we realize the function of using semantics to generate visual features for unseen domains. In addition, we also adopt part of the idea of meta-learning to progressively correct the capability of the generator for better mitigating domain shift problem during the generation process. Through the above ideas, we can detect both seen and unseen bounding boxes and classify them accurately, by combining with the excellent detection ability of Faster-RCNN. Extensive experimental results on two popular datasets, i.e., MSCOCO and KITTI, show that our proposed method can outperform the state-of-the-art methods.

With the rapid development of convolutional neural networks (CNNs), significant progress has been achieved in semantic segmentation. Despite the great success, such deep learning approaches require large scale real-world datasets with pixel-level annotations. However, considering that pixel-level labeling of semantics is extremely laborious, many researchers turn to utilize synthetic data with free annotations. But due to the clear domain gap, the segmentation model trained with the synthetic images tends to perform poorly on the real-world datasets. Unsupervised domain adaptation (UDA) for semantic segmentation recently gains an increasing research attention, which aims at alleviating the domain discrepancy. Existing methods in this scope either simply align features or the outputs across the source and target domains or have to deal with the complex image processing and post-processing problems. In this work, we propose a novel multi-level UDA model named Confidence-and-Refinement Adaptation Model (CRAM), which contains a confidence-aware entropy alignment (CEA) module and a style feature alignment (SFA) module. Through CEA, the adaptation is done locally via adversarial learning in the output space, making the segmentation model pay attention to the high-confident predictions. Furthermore, to enhance the model transfer in the shallow feature space, the SFA module is applied to minimize the appearance gap across domains. Experiments on two challenging UDA benchmarks “GTA5-to-Cityscapes” and “SYNTHIA-to-Cityscapes” demonstrate the effectiveness of CRAM. We achieve comparable performance with the existing state-of-the-art works with advantages in simplicity and convergence speed.

Semantic segmentation and depth estimation play important roles in the field of autonomous driving. In recent years, the advantages of Convolutional Neural Networks (CNNs) have allowed these two topics to flourish. However, people always solve these two tasks separately and rarely solve them in a united model. In this paper, we propose a Mutual Encouragement Network (MENet), which includes a semantic segmentation branch and a disparity regression branch, and simultaneously generates semantic map and visual disparity. In the cost volume construction phase, the depth information is embedded in the semantic segmentation branch to increase contextual understanding. Similarly, the semantic information is also included in the disparity regression branch to generate more accurate disparity. Two branches mutually promote each other during training phase and inference phase. We conducted our method on the popular dataset KITTI, and the experimental results show that our method can outperform the state-of-the-art methods on both visual disparity generation and semantic segmentation. In addition, extensive ablation studies also demonstrate that the two tasks in our method can facilitate each other significantly with the proposed approach.

Summary Cerebral micro‐bleedings are small chronic brain hemorrhages caused by structural abnormalities of the small vessels. CMBs can be found from individuals with stroke at memory clinics and even healthy elderly people. CMBs indicate hemorrhage‐prone pathological states. Research shows that CMBs are associated with an increased risk of future ischemic stroke, intra‐cerebral hemorrhage (ICH), dementia, and death. Considering that CMBs severely influence people's life, it is necessary to identify the CMBs in an early stage to prevent from further deterioration and to help people live a healthy life. In this paper, we proposed using CNN with stochastic pooling for the CMB detection. CNN has good performance in image and video recognition, recommender system, and nature language processing. Based on the collected subject, the experiment result shows that the six‐convolution layer and three fully‐connected layer CNN, nine‐layers in total, achieved sensitivity, specificity, accuracy, and precision as 97.22%, and 97.35%, 97.28%, and 97.35% in average of ten runs, which shows better performance than five state‐of‐the‐art methods.