This study proposes an Efficient and Explainable Ensemble-learning framework (EEE-framework) designed for early detection of non-small cell lung cancer (NSCLC) biomarkers using gene expression data, specifically addressing challenges of interpretability in detection outcomes. The EEE-framework comprises two core modules. The first module constructs an ensemble-learning model balancing detection accuracy and interpretability. Considering the typical trade-off between accuracy and interpretability, we explore various combinations of 10 individual learners, ultimately selecting five (Decision Tree, Random Forest, XGBoost, AdaBoost, and Gaussian Naive Bayes) with high interpretability as base learners. Subsequently, we classify NSCLC by integrating voting-ensemble with high interpretability. The second module develops a multi-perspective approach to identify important NSCLC biomarkers using six explainable artificial intelligence (XAI) methods, ranging from Local to Global Interpretability and from Intrinsic to Post-hoc Interpretability. The EEE-framework's generalization ability and interpretability are validated using the TCGA RNA_seq public dataset and a self-constructed methylation dataset. Validation results demonstrate the exceptional classification performance of the ensemble-learning model integrated in the framework, achieving F1 values of 0.9983(AUC:0.9993) and 0.8462(AUC:0.9249) on the two datasets, respectively. Global and Local Interpretable Visualization results significantly enhance the diagnosis and understanding of NSCLC biomarkers, providing insights to their importance rankings, interrelationships, and causality with predicted outcomes.

Most fine-grained visual recognition methods endeavor to directly locate discriminative regions in intricate environments, but tend to overlook the object’s holistic structure, which may lead to misclassification due to overemphasizing incorrect areas. In this paper, we propose a coarse-to-fine paradigm, which prioritizes locating holistic structural regions of the target object, followed by a gradual search to locate discriminative areas. Specifically, we first design the “look into object” module to locate the areas encompassing the target’s holistic structure using prior information. Subsequently, without introducing additional parameters, we design a partial focus searching module to enhance feature representations of discriminative regions within the target’s structural composition. Ultimately, we segregate the foreground components from the original image, attaining a more precise characterization of the target. Furthermore, we demonstrate the practical application potential of our model in real-world industries through our self-constructed DHU-Fine-grained-6000 dataset. Comparative experiments on three public datasets indicate that the superiority of our approach over many recent methods and holds promising application potential in industrial production processes.

Multivariate time series prediction is a significant research area that aims to forecast future values based on past observations. Deep learning models with attention mechanisms have shown good predictive performance by emphasizing optimal-related sequences in the target series. However, these models ignore mutation information of nontarget sequences and the long short-term dependencies. To this end, a divide-and-rule combined learning method is proposed to address these limitations, which uses differentiated feature extractors to process different implicit features. First, we design a spatial and temporal information extractor to extract the time-dimensional feature information in the separation stage. Then, a multivariate mutation information extractor is constructed by convolution and maximum pooling layer to capture mutation information of nontarget sequences. Subsequently, the decoder component of the encoder-decoder model extracts long short-term dependencies while preserving the information of the target sequence to be predicted. Finally, in the cooperation stage, a feature fusion method based on a point attention mechanism is proposed, which can assign individual weights to each feature point and enhance the ability to focus on local areas. Experimental results on five real datasets in different domains show that the proposed method has better predictive performance compared to other baseline models.

Swarm intelligence optimization algorithms (SIOAs) are widely used to address complex problems but often trapped in local optima. To overcome this, we propose a novel multi-population SIOA based on cellular coordination mechanisms in immune systems (referred as CCOA), inspired by the immune system's efficient elimination of viruses. The CCOA consists of four key units: an intra-population cell communication unit, an inter-population cell communication unit, a cell migration unit, and a cell division unit. Through collaborative communication, the intra-population and inter-population units guide cells towards global optima, enhancing the algorithm's global search ability. The cell migration unit improves convergence speed by guiding cell movements. The cell division unit allows cells to divide in large numbers during the late iterations for improved convergence accuracy. We evaluate the effectiveness of the CCOA using eight standard test functions and apply it to river flow prediction in an Elman neural network (referred as CCOA-Elman). The results obtained from optimizing test functions demonstrate that the CCOA outperforms existing algorithms in accuracy, convergence, and stability. The real-world application illustrates that the CCOA-Elman achieves over 95% prediction accuracy, surpassing the accuracies of three other models. The CCOA offers a promising approach to overcome local optima in optimization.

In recent years, neural architecture search (NAS) methods have been proposed for the automatic generation of task-oriented network architecture in image classification. However, the architectures obtained by existing NAS approaches are optimized only for classification performance and do not adapt to devices with limited computational resources. To address this challenge, we propose a neural network architecture search algorithm aiming to simultaneously improve the network performance and reduce the network complexity. The proposed framework automatically builds the network architecture at two stages: block-level search and network-level search. At the stage of block-level search, a gradient-based relaxation method is proposed, using an enhanced gradient to design high-performance and low-complexity blocks. At the stage of network-level search, an evolutionary multiobjective algorithm is utilized to complete the automatic design from blocks to the target network. The experimental results demonstrate that our method outperforms all evaluated hand-crafted networks in image classification, with an error rate of 3.18% on Canadian Institute for Advanced Research (CIFAR10) and an error rate of 19.16% on CIFAR100, both at network parameter size less than 1 M. Obviously, compared with other NAS methods, our method offers a tremendous reduction in designed network architecture parameters.

Particle swarm optimization (PSO) tends to fall into local optimum during the high-dimensional optimization process To address this limitation, a hybrid optimization approach by combining PSO with mechanisms in neuro-endocrine-immune systems (NEI-PSO) is proposed. The NEI-PSO includes a nervous guidance unit, an endocrine regulation unit, and an immune orientation unit. The nervous guidance unit and the immune orientation unit are designed based on the nervous system guidance mechanism and the immune system orientation mechanism respectively. Through the joint effect of these two units, the update mode of particle movement is changed; as a result, the global search ability of the NEI-PSO can be improved. The endocrine regulation unit changes the learning factor based on the hormone regulation law of the endocrine system, and in turn improves the optimization convergence speed of the proposed approach. In this paper, the NEI-PSO is evaluated using eight high-dimensional benchmark functions and a real-world high-dimensional optimization application for a non-Pieper six-axis robot. The results demonstrate that the proposed NEI-PSO approach has prominent advantages in search accuracy, convergence ability, and stability, compared to some existing optimization approaches.

In multi-camera video tracking, the tracking scene and tracking-target appearance can become complex, and current tracking methods use entirely different databases and evaluation criteria. Herein, for the first time to our knowledge, we present a universally applicable template library updating approach for multi-camera human tracking called multi-state self-learning template library updating (RS-TLU), which can be applied in different multi-camera tracking algorithms. In RS-TLU, self-learning divides tracking results into three states, namely steady state, gradually changing state, and suddenly changing state, by using the similarity of objects with historical templates and instantaneous templates because every state requires a different decision strategy. Subsequently, the tracking results for each state are judged and learned with motion and occlusion information. Finally, the correct template is chosen in the robust template library. We investigate the effectiveness of the proposed method using three databases and 42 test videos, and calculate the number of false positives, false matches, and missing tracking targets. Experimental results demonstrate that, in comparison with the state-of-the-art algorithms for 15 complex scenes, our RS-TLU approach effectively improves the number of correct target templates and reduces the number of similar templates and error templates in the template library.

The management of recreational fishing is a controversial subject in most jurisdictions. On the one hand, recreational fishing provides substantial economic benefits. Fishing activities, on the other hand, can threaten valuable fish stocks and cause damage to marine environments. However, most management strategies undertaken so far tend to be ad hoc and not supported by proper modeling. To address this, this paper proposes an integrated agent-based model of recreational fishing and a coral reef environment as a platform to evaluate both the economic and biophysical impacts. The platform is used to evaluate the effects of an incentive-based scheme for the management of recreational fishing in coral reef environment. A negative incentive in the form of fish extraction fee is evaluated. Simulation of fish extraction fees indicate that high value fish biomass can be increased substantially when fish extraction fees are pushed to a certain critical but narrow range. Anglers benefit from high catch rates and this compensation dampens the degree of economic welfare loss among anglers. The results from the simulations demonstrate the extent to which the often controversial subject of recreational fishing management can be facilitated using integrated modeling.