Computational Intelligence最新文献

Age-related macular degeneration (AMD) is a progressive retinal disease that can lead to vision loss if not diagnosed early. Accurate and timely detection is essential for effective treatment. This study presents a comprehensive deep learning-based framework for the automated detection of AMD using optical coherence tomography (OCT) images. The proposed method integrates advanced preprocessing, segmentation, feature extraction, and classification techniques. An improved Wiener filter is applied to enhance image quality by reducing noise while preserving edge details. A U-Net architecture is then used for accurate segmentation of retinal regions. Subsequently, diverse features such as histogram of oriented gradients (HOG), gray-level co-occurrence matrix (GLCM), scale-invariant feature transform (SIFT), VGG16, statistical features, and improved median ternary pattern (MTP) are extracted from segmented images. Then, an enhanced principal component analysis (PCA) with feature scaling is used for the optimal selection of features to improve classification robustness. Finally, a hybrid model combining deep convolutional neural networks (DCNNs) and AlexNet is proposed for AMD classification. Moreover, the proposed model is evaluated against traditional methods and achieves superior performance, with an accuracy of 0.958 and F-measure of 0.964. Experimental results confirm the model's effectiveness and reliability, highlighting its potential for clinical application in early AMD detection and diagnosis.

Multimodal data can more vividly and strangely represent consumers' feelings and sentiments than single-modal content. Users are moving beyond traditional text-based content on social media and are now more commonly incorporating images and text to convey their experiences and articulate their opinions. Traditional text-based techniques have given way to multimodal sentiment analysis, which poses a more complex challenge. This work attempts to address the challenge of sentiment analysis in image-text posts by presenting a refined recognition method that efficiently leverages textual and visual content information. Initially, a single dataset can be created by combining the MVSA-Single and MVSA-Multiple datasets. It is gathered to improve data quality by pre-processing the data with a Gaussian bilateral filter (GBF) for picture features and lemmatization, stemming, and stop word removal for text characteristics. The normalized term-inverse document frequency (NorTID) model is used to extract text features. The Convolutional VGG-16 (ConV-16) model extracts image characteristics. The Dense Stacked Long Short-Term Memory Network (DStaL) model is used to independently analyze the collected multimodal information. The obtained features are fused together, and the sentiments are effectively classified using the Multilevel Fusion Classifier (MFuse) model. The study illustrates the enhanced efficacy of the suggested strategy by comparing the results to those of standard approaches using several performance measures. For the MVSA-Single and MVSA-Multiple datasets, the findings show a 97.3% accuracy, 94.8% precision, 94.8% recall, 97.3% specificity, 92.1% Matthews Correlation Coefficient (MCC), 36.3% Root Mean Squared Error (RMSE), and 0.07% Mean Absolute Error (MAE).

For solving the problem of dynamic task topology change in multi-manipulator cooperative system, a distributed control scheme based on topology transformation is proposed. In this scheme, a topological transformation mechanism is introduced so that the multi-manipulator cooperative system dynamically adapts to the multi-stage tasks with topological changes. To facilitate the solution of the problem, the distributed scheme is converted into an optimization problem. In addition, a gradient neural network solver with an activation function is designed, which greatly improves the convergence speed of the solver. Theoretical analysis is given, which confirms the convergence of the solver. Furthermore, the effectiveness of the proposed scheme is demonstrated.

With the rapid expansion of intelligent medical equipment and their interconnectedness through the Internet of Things (IoT), addressing safety issues in the communicating system has become increasingly critical. A learning mechanism is proposed for an intelligent healthcare-based communication system that uses blockchain for secure network communication and incorporates a data evaluation layer based on cloud which actively segregates and ranks transactions into three main categories: Good, Moderate, and Malware. Fog servers are utilized to route the communicating nodes via Rician and Rayleigh channels. The learning mechanism employs a deep neural network to instruct and classify categories, thereby improving the blockchain layer's decision-making process. This paper introduces several significant contributions, such as the development of a secure blockchain framework for user authentication and a protected digital ledger for communication. Additionally, it incorporates a cloud-driven data analysis layer combined with a neural network to improve training accuracy and category classification. The developed algorithm surpassed the existing works in terms of quality of service (QoS) parameters with low latency, bit error rate (BER), higher signal to inference plus noise ratio (SINR), packet delivery ratio (PDR), true detection rate (TDR), false detection rate (FDR), and throughput. Also, a thorough comparison of consensus mechanisms like practical Byzantine fault tolerance (pBFT), proof of work (PoW), Raft, and Paxos is done to ensure which consensus helps optimize the proposed system in terms of security and fault tolerance with low latency and energy-efficient operations. It also establishes a secure and efficient communication network for smart healthcare, aimed at enhancing the overall quality of life for individuals.

Current diagnosis of autism spectrum disorder (ASD) and developmental language/reading disorder (DLD/DD) relies predominantly on subjective behavioral assessments, this underscoring the urgent need for objective biomarkers to enable early intervention. This study proposes a Multi-Tier Transformer (MTT) model for early identification of ASD and DLD/DD using resting-state EEG baseline power values. To address severe class imbalance, we augmented the dataset using the SMOTE method. The MTT architecture integrates a Feature-Embedding layer, a Feature-Attention mechanism that dynamically weights multi-spectral inputs, and a dual-attention encoding block comprising both self-attention and cross-attention to enhance contextual representation learning from limited samples. Transfer learning was further employed to improve robustness by pre-training on augmented data and fine-tuning on original samples. Evaluated on clinical infant EEG data, the proposed MTT achieved an accuracy of 0.91 (95% CI: 0.89–0.93), recall of 0.89, and AUC of 0.97, significantly outperforming the state-of-the-art FT-transformer (p = 0.00091). The results indicate that MTT provides a robust and interpretable deep learning tool for auxiliary diagnosis of neurodevelopmental disorders in infancy.

Personalized federated learning represents a pivotal strategy for addressing the challenges posed by statistical heterogeneity in federated learning. Clients optimize their models by leveraging information from other clients through a global model. However, data heterogeneity constrains the generalization capacity of the global model, thereby degrading the feature representation capability of local client models, especially in clients with limited data. In response to this challenge, we propose the Federal Merging Subgroup Information (FedMSI) method to augment personalized information in personalized federated learning. On the server side, FedMSI employs model clustering to identify subgroups of clients with similar personalized data distributions. It then aggregates cluster center models within each subgroup and transmits them to clients for use in the subsequent round of assisted training. On the client side, FedMSI introduces a local optimization objective that incorporates the cluster center model, enabling the extraction of informative knowledge to enhance local training. Experiments demonstrate the effectiveness of FedMSI across different datasets, data heterogeneity levels, and data sizes. Ablation experiments further confirm the effectiveness of the design of the local optimization objective. Compared to state-of-the-art methods, FedMSI achieves a 13.16% improvement in scalability performance accuracy.

As modelling of nonlinear oscillator systems plays an important part in science and engineering fields, a double loop Radial Basis Function Neural Network (RBFNN) with adaptive radius and lattices is proposed for handling this issue. In this design, a large enough lattice arranged to cover all of the trajectories is taken as the mapping center of the RBFNN at the initial condition. The number of lattices will be dynamically adjusted, and those lattices far from the trajectories will be removed. Applying Taylor expansion in local space, the activated radius factor can be separated from the Gaussian function. In order to guarantee that the modelling scheme has the characteristic of fast convergence, the error power function is utilized to minimize the gain parameter of the error differential equation. In the double loop structure, the updated equation of weights and activated radius can be determined by the Lyapunov function, which can guarantee that the weights and the activated radius will converge to the neighborhood of their true value and the tracking error of state trajectories will converge to the neighborhood of zero. In order to show the effectiveness and superiority of the double loop RBFNN proposed in this paper, Helmholtz–Duffing and Vanderpol–Duffing are used as the testing objects of the nonlinear oscillator system while comparing with Deterministic Learning.

Aspect-based Sentiment Analysis (ABSA) is a vital NLP task that identifies sentiment towards specific entities or aspect terms within a text. Recently, large language models (LLMs) have shown impressive capabilities in semantic comprehension and logical inference. However, LLM hallucinations pose challenges in accurately determining sentiment polarity for aspect terms, leading to performance issues. Moreover, current ABSA methods often fail to fully leverage the vast prior knowledge embedded within LLMs, resulting in suboptimal classification outcomes for specific aspects. Inspired by these challenges, we propose the BYD-OBS-ABSA framework—‘Beyond Simple Observations, Embracing Comprehensive Contextual Insights’ for ABSA tasks. This framework leverages unique in-context constraints, backgrounds, and analogical reasoning to address LLM hallucinations and uses self-adaptive bootstrap instructions optimization to enhance LLM predictions. BYD-OBS-ABSA integrates various in-context augmentation strategies, including emotion-oriented backgrounds, constraints, and analogical reasoning. BYD-OBS-ABSA further improves initial LLM instructions through adaptive iterative optimization using a random search bootstrap algorithm, maximizing the benefits of LLM prompting. Extensive zero/few-shot experiments with GPT-3.5-turbo across six public datasets validate the effectiveness and robustness of our framework, even surpassing human judgment in certain scenarios.