CAAI Transactions on Intelligence Technology最新文献

This paper introduces a novel lightweight colour image encryption algorithm, specifically designed for resource-constrained environments such as Internet of Things (IoT) devices. As IoT systems become increasingly prevalent, secure and efficient data transmission becomes crucial. The proposed algorithm addresses this need by offering a robust yet resource-efficient solution for image encryption. Traditional image encryption relies on confusion and diffusion steps. These stages are generally implemented linearly, but this work introduces a new RSP (Random Strip Peeling) algorithm for the confusion step, which disrupts linearity in the lightweight category by using two different sequences generated by the 1D Tent Map with varying initial conditions. The diffusion stage then employs an XOR matrix generated by the Logistic Map. Different evaluation metrics, such as entropy analysis, key sensitivity, statistical and differential attacks resistance, and robustness analysis demonstrate the proposed algorithm's lightweight, robust, and efficient. The proposed encryption scheme achieved average metric values of 99.6056 for NPCR, 33.4397 for UACI, and 7.9914 for information entropy in the SIPI image dataset. It also exhibits a time complexity of $��O��\left(��2�\times �M�\times �N��\right)��$ for an image of size $��M�\times �N�$.

Using sarcasm on social media platforms to express negative opinions towards a person or object has become increasingly common. However, detecting sarcasm in various forms of communication can be difficult due to conflicting sentiments. In this paper, we introduce a contrasting sentiment-based model for multimodal sarcasm detection (CS4MSD), which identifies inconsistent emotions by leveraging the CLIP knowledge module to produce sentiment features in both text and image. Then, five external sentiments are introduced to prompt the model learning sentimental preferences among modalities. Furthermore, we highlight the importance of verbal descriptions embedded in illustrations and incorporate additional knowledge-sharing modules to fuse such image-like features. Experimental results demonstrate that our model achieves state-of-the-art performance on the public multimodal sarcasm dataset.

Most prevailing attention mechanism modules in contemporary research are convolution-based modules, and while these modules contribute to enhancing the accuracy of deep learning networks in visual tasks, they concurrently augment the overall model complexity. To address the problem, this paper proposes a plug-and-play algorithm that does not increase the complexity of the model, Laplacian attention (LA). The LA algorithm first calculates the similarity distance between feature points in the feature space and feature channel and constructs the residual Laplacian matrix between feature points through the similarity distance and Gaussian kernel. This construction serves to segregate non-similar feature points while aggregating those with similarities. Ultimately, the LA algorithm allocates the outputs of the feature channel and the feature space adaptively to derive the final LA outputs. Crucially, the LA algorithm is confined to the forward computation process and does not involve backpropagation or any parameter learning. The LA algorithm undergoes comprehensive experimentation on three distinct datasets—namely Cifar-10, miniImageNet, and Pascal VOC 2012. The experimental results demonstrate that, compared with the advanced attention mechanism modules in recent years, such as SENet, CBAM, ECANet, coordinate attention, and triplet attention, the LA algorithm exhibits superior performance across image classification, object detection and semantic segmentation tasks.

Intelligent sorting is an important prerequisite for the full quantitative consumption and harmless disposal of kitchen waste. The existing object detection method based on an ImageNet pre-trained model is an effective way of sorting. Owing to significant domain gaps between natural images and kitchen waste images, it is difficult to reflect the characteristics of diverse scales and dense distribution in kitchen waste based on an ImageNet pre-trained model, leading to poor generalisation. In this article, the authors propose the first pre-trained model for kitchen waste sorting called KitWaSor, which combines both contrastive learning (CL) and masked image modelling (MIM) through self-supervised learning (SSL). First, to address the issue of diverse scales, the authors propose a mixed masking strategy by introducing an incomplete masking branch based on the original random masking branch. It prevents the complete loss of small-scale objects while avoiding excessive leakage of large-scale object pixels. Second, to address the issue of dense distribution, the authors introduce semantic consistency constraints on the basis of the mixed masking strategy. That is, object semantic reasoning is performed through semantic consistency constraints to compensate for the lack of contextual information. To train KitWaSor, the authors construct the first million-level kitchen waste dataset across seasonal and regional distributions, named KWD-Million. Extensive experiments show that KitWaSor achieves state-of-the-art (SOTA) performance on the two most relevant downstream tasks for kitchen waste sorting (i.e. image classification and object detection), demonstrating the effectiveness of the proposed KitWaSor.

Audio-visual scene classification (AVSC) poses a formidable challenge owing to the intricate spatial-temporal relationships exhibited by audio-visual signals, coupled with the complex spatial patterns of objects and textures found in visual images. The focus of recent studies has predominantly revolved around extracting features from diverse neural network structures, inadvertently neglecting the acquisition of semantically meaningful regions and crucial components within audio-visual data. The authors present a feature pyramid attention network (FPANet) for audio-visual scene understanding, which extracts semantically significant characteristics from audio-visual data. The authors’ approach builds multi-scale hierarchical features of sound spectrograms and visual images using a feature pyramid representation and localises the semantically relevant regions with a feature pyramid attention module (FPAM). A dimension alignment (DA) strategy is employed to align feature maps from multiple layers, a pyramid spatial attention (PSA) to spatially locate essential regions, and a pyramid channel attention (PCA) to pinpoint significant temporal frames. Experiments on visual scene classification (VSC), audio scene classification (ASC), and AVSC tasks demonstrate that FPANet achieves performance on par with state-of-the-art (SOTA) approaches, with a 95.9 F1-score on the ADVANCE dataset and a relative improvement of 28.8%. Visualisation results show that FPANet can prioritise semantically meaningful areas in audio-visual signals.

Yong Li, Shuhang Wang, Shijie Xu, and Jiao Yin. 2024. Trustworthy semi-supervised anomaly detection for online-to-offline logistics business in merchant identification. CAAI Transactions on Intelligence Technology 9, 3 (June 2024), 544–556. https://doi.org/10.1049/cit2.12301.

In the section discussing the spatial distribution of fraud and normal merchants' shipping addresses, the following text needs correction:

Please replace Figure 1 and 2 with the following text ‘According to the data analysis results, the spatial distribution of fraud merchants' shipping addresses is characterised by sparsity (because fraud merchants ship on behalf of others, resulting in a large number of shipping addresses with few shipments per address), while the distribution of normal merchants' shipping addresses is characterised by density (as normal merchants typically ship from centralised warehouses, resulting in a small number of shipping addresses with a large number of shipments per address). These differences in shipping behaviour can provide significant assistance in detecting fraud merchants.’

We apologise for this error.

Please note that due to the deletion of two images, the order of subsequent images has been adjusted accordingly.

Currently, the link prediction algorithms primarily focus on studying the interaction between nodes based on chain structure and star structure, which predominantly rely on low-order structural information and do not explore the multivariate interactions between nodes from the perspective of higher-order structural information present in the network. The cycle structure is a higher-order structure that lies between the star and clique structures, where all nodes within the same cycle can interact with each other, even in the absence of direct edges. If a node is encompassed by multiple cycles, it indicates that the node interacts and associates with a greater number of nodes in the network, and it means the node is more important in the network to some extent. Furthermore, if two nodes are included in multiple cycles, it signifies the two nodes are more likely to be connected. Therefore, firstly, a multi-information fusion node importance algorithm based on the cycle structure information is proposed, which integrates both high-order and low-order structural information. Secondly, the obtained integrated structure information and node feature information is regarded as the input features, a two-channel graph neural network model is designed to learn the cycle structure information. Then, the cycle structure information is utilised for the task of link prediction, and a graph neural link predictor with multi-information interactions based on the cycle structure is developed. Finally, extensive experimental validation and analysis show that the node ranking result of the proposed node importance index is more consistent with the actual situation, the proposed graph neural network model can effectively learn the cycle structure information, and using higher-order structural information—cycle information proves to significantly enhance the overall link prediction performance.

This paper focuses on the problem of multi-station multi-robot spot welding task assignment, and proposes a deep reinforcement learning (DRL) framework, which is made up of a public graph attention network and independent policy networks. The graph of welding spots distribution is encoded using the graph attention network. Independent policy networks with attention mechanism as a decoder can handle the encoded graph and decide to assign robots to different tasks. The policy network is used to convert the large scale welding spots allocation problem to multiple small scale single-robot welding path planning problems, and the path planning problem is quickly solved through existing methods. Then, the model is trained through reinforcement learning. In addition, the task balancing method is used to allocate tasks to multiple stations. The proposed algorithm is compared with classical algorithms, and the results show that the algorithm based on DRL can produce higher quality solutions.

The bat algorithm (BA) is a metaheuristic algorithm for global optimisation that simulates the echolocation behaviour of bats with varying pulse rates of emission and loudness, which can be used to find the globally optimal solutions for various optimisation problems. Knowing the recent criticises of the originality of equations, the principle of BA is concise and easy to implement, and its mathematical structure can be seen as a hybrid particle swarm with simulated annealing. In this research, the authors focus on the performance optimisation of BA as a solver rather than discussing its originality issues. In terms of operation effect, BA has an acceptable convergence speed. However, due to the low proportion of time used to explore the search space, it is easy to converge prematurely and fall into the local optima. The authors propose an adaptive multi-stage bat algorithm (AMSBA). By tuning the algorithm's focus at three different stages of the search process, AMSBA can achieve a better balance between exploration and exploitation and improve its exploration ability by enhancing its performance in escaping local optima as well as maintaining a certain convergence speed. Therefore, AMSBA can achieve solutions with better quality. A convergence analysis was conducted to demonstrate the global convergence of AMSBA. The authors also perform simulation experiments on 30 benchmark functions from IEEE CEC 2017 as the objective functions and compare AMSBA with some original and improved swarm-based algorithms. The results verify the effectiveness and superiority of AMSBA. AMSBA is also compared with eight representative optimisation algorithms on 10 benchmark functions derived from IEEE CEC 2020, while this experiment is carried out on five different dimensions of the objective functions respectively. A balance and diversity analysis was performed on AMSBA to demonstrate its improvement over the original BA in terms of balance. AMSBA was also applied to the multi-threshold image segmentation of Citrus Macular disease, which is a bacterial infection that causes lesions on citrus trees. The segmentation results were analysed by comparing each comparative algorithm's peak signal-to-noise ratio, structural similarity index and feature similarity index. The results show that the proposed BA-based algorithm has apparent advantages, and it can effectively segment the disease spots from citrus leaves when the segmentation threshold is at a low level. Based on a comprehensive study, the authors think the proposed optimiser has mitigated the main drawbacks of the BA, and it can be utilised as an effective optimisation tool.