ABSTRACT Diabetic retinopathy (DR) and diabetic macular edema (DME) are eye‐related diseases most commonly seen in persons with diabetes. Existing models did not achieve the desired level of accuracy or success in distinguishing DR and DME. Thus, by resolving this issue, the performance and accuracy of automated classification systems are significantly improved. Therefore, a transfer learning model is developed to simultaneously classify DR and DME grades in retinal images. Initially, the input retinal fundus images are collected from standard databases. These images are given to the proposed Transfer Learning‐based Adaptive Fused Convolutional Neural Networks (TL‐AFCNN). This model fusion is created by using a Residual Attention Network (RAN) with multidilated DenseNet techniques. In order to improve the effectiveness of simultaneous classification using TL‐ACNN, the weights from this network are fine‐tuned via the Fitness Oriented Lotus Effect Optimization Algorithm (FO‐LEA). From, the developed TL‐ACNN model, simultaneous grade classification outcomes of both DR and DME are attained. Therefore, the risk of vision loss is treated effectively by analyzing the grade‐classified outcomes from the proposed model. Finally, the simultaneous classification results attained by the suggested model are compared with existing models to verify its potential.
ABSTRACT The revenue generated from online ads has become quite significant, and as with the advancement in any sort of business, this one brings fraudsters with it. However, different forms of fraud can be performed as this work tackles the problems of click fraud in advertisements. In this case, the fraud can be performed by the party that bought the advertisement to boost its revenue, or by other malicious parties that tend to exhaust the resources for the said ad, for example. Due to this and many other scenarios, a robust solution for detecting such cases must be established. However, existing click fraud detection approaches either rely on static rule‐based systems or deep learning models with manually tuned hyperparameters, which may result in limited adaptability and suboptimal performance in complex sequential environments. Therefore, there is a need for an adaptive optimization strategy capable of effectively tuning sequential models for improved fraud detection accuracy. This work proposes three different types of recurrent neural networks (RNNs) that are combined with the attention mechanism. Furthermore, in each of the three different experiments, the networks were optimized by strong metaheuristics optimizers, the results of which were compared to establish the strongest one. This was done with the purpose of confirming the improvements to the variable neighborhood search (VNS) algorithm, which was proposed by the authors in this work. The best synthesized RNN model tuned by the suggested modified optimizer attained accuracy of 0.806569, with Matthews correlation coefficient of 0.613209.
This research expands upon our previous work on few-shot learning for multiclass classification on tabular data. Expanding on the FSL-LFMG framework, which integrates latent feature augmentation, advanced STUNT multitasking generation, and Prototypical Networks, we aim to illustrate the architecture’s flexibility in adapting to new concept tasks. Our prior research demonstrated the efficacy of FSL-LFMG in few-shot multiclass classification, outperforming traditional Tree Ensemble models and One-vs-the-rest classifiers. This research further investigates the framework’s generalization capabilities by applying it to a new concept tasks, specifically predicting customer loyalty levels based on an existing NPS segmentation model. Experimental results demonstrate substantial performance in both 1-shot and 5-shot configurations, attaining mean test accuracies of 83.95
In recent years, Deep Learning (DL) models have shown excellent results in various fields, including healthcare, but their “opaque” nature limits their trust and adoption in critical contexts. In this work, we present XAIMed2.0, an application designed to support healthcare professionals in interpreting the decisions of CNN models applied to medical image analysis. The application integrates several eXplainable AI (XAI) techniques to offer different visual explanations. The application integrates multiple eXplainable AI (XAI) techniques to generate visual explanations. Specifically, it employs several saliency map methods (Grad-CAM, Grad-CAM++, RISE, and LIME) and merges them into a cumulative visualization that highlights the regions of highest consensus across methods. To complement the visual explanations, the system uses a Large Language Model (LLM) to generate textual explanations in natural language, capable of contextualizing the results, highlighting discrepancies, and improving clinician understanding.
Pre-trained computer vision models are readily available and may easily be downloaded and applied to new problems using transfer learning. The main barrier to wider uptake of computer vision models is now the availability of annotated training data, particularly for specialist tasks for which no annotated set exists. The human process of hand-annotating images is expensive, slow, and error-prone. This paper presents methods for automatically generating annotated training data for computer vision models, including generating synthetic data using computer games engine software and bootstrapping large annotation sets from a small number of manual annotations. The role of image augmentation is also explored, and we show how to produce life-like augmentations in synthetic data and how augmentation improves automated annotation techniques. We call this process Image Engineering and present experimental results and illustrative case studies.
Money laundering is the process of disguising illicit funds (“dirty money”) as legitimate assets. As the third-largest criminal industry globally, it poses significant threats to economic stability and national security. To counter this, financial institutions implement anti-money laundering (AML) frameworks, which involve customer identity verification and risk assessment, transaction monitoring to detect suspicious activity, and reporting suspicious transactions to regulators. This article focuses on the transaction monitoring component of AML. We propose a Graph Convolutional Network (GCN) model enhanced with node2vec embeddings to classify transactions as legitimate or suspicious. Unlike prior work, our model is fine-tuned using evaluation objectives aligned with current industry practice. Furthermore, we experimentally demonstrate how the class imbalance problem, which is typical of AML data, adversely impacts classification performance. We present comprehensive experiments to demonstrate and justify how important parameters of the model were optimized and selected, enabling reproducibility and future benchmarking. We also discuss how the proposed model can rank transaction alerts by offense likelihood to expedite manual investigations. Finally, experimental results show that the proposed GCN model outperforms traditional machine learning techniques and prior works that used the same dataset.
Interpreting the human visual system as a cascading sequence of an information system blocks, we, relying on the experience of developing information technologies (data encoding/decoding algorithms) and using specialized neurophysiological knowledge, have developed a model of an neuromorphic system, which in turn maximally imitates the perceptual mechanisms of the visual system. Essentially, we develop a neuromorphic system for processing data streams (registered counts) by extending Marr’s primal sketch paradigm with encoder blocks and implementing it at the algorithmic level. This article presents the results of our previous work in a more detailed and substantiated manner, as well as some new results. The results include a number of data representations internal to the neuromorphic system and algorithmic procedures designed to process them. Most of the obtained theoretical results are illustrated by the results of computer simulation.
The Traveling Salesperson Problem (TSP) is maybe one of the most important 𝒩𝒫 -complete optimization problems. Due to its many local optima, simple local searches like the (1+1) EA or the Randomized Local Search (RLS) are not suitable to solve the TSP. Our past research has shown that plugging Frequency Fitness Assignment (FFA) into such local searches can significantly improve their chance to reach the global optima. We showed that FRLS, i.e., RLS with FFA plugged-in, can find the optima of many more instances of the TSP than pure local search. In these prior works, we concluded that the reason for this good performance probably is ability of FFA to escape from local optima. However, truly exploring how FFA impacts the search is complicated. For most of the available TSP benchmark instances, the number #L of local optima, the sizes BL of their basins of attraction, as well as the overall number M of different tour lengths are unknown. We believe that these features will have a strong impact on the performance of RLS, FRLS, as well as all other heuristics for the TSP. Yet, determining them is not feasible either, because this would be even much more complicated than solving the instances to optimality. Knowing M , for example, does not just require to finding the globally optimal solution, but one solution corresponding at each and every possible tour length. On any problem of non-trivial size an structure, this is impossible. In this work, we pursue another route: We generate new small-scale instances for which the exact values of these parameters can be known. This allows us to explore the performance of both RLS and FRLS in settings where we know all problem instance parameters that could possibly affect the algorithms. We can therefore, for the first time on the TSP, clearly observe and confirm the ability of FRLS to escape from local optima and we can support the claim that M has a substantial impact on its performance with very conclusive evidence.
Deep learning and computer-aided detection (CAD) systems are increasingly vital for the early detection of lung cancer, where timely diagnosis can substantially improve patient survival. In this study, we present a two-phase deep learning framework for automated lung nodule segmentation using the Luna-16 dataset. In the first phase, the LungQuant algorithm was employed to segment the lung parenchyma, achieving an average Dice Similarity Coefficient (DSC) of 90
Evolution relies fundamentally upon two intertwined processes: speciation, where new reproductively isolated species arise from ancestral ones, and subsequent adaptation, where these isolated species evolve increasingly complex traits. This research presents an environment for simulating both of these dynamics using a basic genetic algorithm and simple simulated agents. First, we show that allopatric speciation is achievable without a priori knowledge of the environment. In our simulations, agents utilize only secondary, fitness-neutral characteristics to guide reproductive choices, creating reproductively isolated species without direct fitness comparison. This method led to a more complete separation of species than previously achieved. In addition to simulating allopatric speciation, we show that a simple genetic algorithm is capable of evolving complex features despite detrimental intermediate mutations. Our simulations demonstrate that the evolution of complex traits is feasible, though highly sensitive to the number of detrimental mutations required before such traits confer a fitness advantage. We then explore the effects of a combined simulation, allowing both speciation and complex trait evolution to take place, resulting in natural selection and competition between species without explicit fitness information.
Landscape analysis and automated algorithm selection and tuning are the subject of intense ongoing research in multi-objective optimization. Landscape analysis captures the characteristics of a problem’s landscape through numerical features. This provides machine learning models with valuable information to make appropriate decision. In addition to landscape analysis, it is also important to quantify algorithms’ component characteristics, independently of the characteristics of the problem being solved, in order to analyze and understand their relationship to performance. One of such key components is crossover. Crossover has been the focus of numerous studies. Its effectiveness depends on the operator’s properties to mix information, the specific characteristics of the problem, and the diversity of the population, influenced by the dynamics of the algorithm. This study focuses on binary representations and introduces a method to examine the relationship between crossover features and the performance of a multi-objective evolutionary algorithm on problem subclasses with random and neighbor patterns of variable interactions. The aim is to identify the crossover features relevant to performance in each problem subclass through regression models.
In this work, we propose an innovative anomaly detection platform, called METAL that integrates Multimodal-feature Extraction (ME) and Transformer-based AutoEncoder (TA) with a Location (L) classifier. Our contributions to this work are six-fold: (1) develop a signal averaging method to remove unrelated noise from the raw sensor data to improve the quality of signals; (2) extract the meaningful signal properties from three vibration signal directions using ME methods, fuse these properties together, and then reduce data dimensionality using PCA to generate a refined PCA feature set; (3) enhance a TA model to learn pump behavior from the extracted PCA feature set to detect anomalous behavior; (4) integrate a location classifier into our enhanced TA model in the training process to highlight distinctions between locations and provide feedback for enhancing performance within the AutoEncoder framework; (5) conduct an extensive experimental case study on the Stonecutters Island Sewage Treatment Works located in Hong Kong, in which we reconstruct the vibration signals of the pumps and set a threshold to detect anomalies based on reconstruction error. Having labeled data on the dates when the pumps were operating abnormally, we are able to obtain the accuracy metrics on the anomaly detection performance of the METAL pipeline. METAL achieves MCC/F1 scores of 0.961/0.972 on 1 location, 0.925/0.936 on 2 locations, and 0.768/0.821 on 6 locations that outperform the existing approaches; and (6) design and implement a web-based prototype for a Sewage Pump Monitoring System hosting the METAL pipeline, providing an interactive interface for future use.
ABSTRACT Breast cancer (BC) is the most common disease among women globally and a major contributor to both illness and death. The timely and precise detection and classification (DAC) of this life‐threatening disease are vital in reducing mortality rates and preventing further complications. Traditional BC detection methods are often time‐consuming and costly. This research proposes a Deep Explainable Breast Cancer Detection and Classification Network (DeepEBCDCNet) to enable precise and early diagnosis of BC using histopathological images (HI). The DeepEBCDCNet model consists of 13 learnable layers, including nine convolutional layers (CN) followed by four fully connected (FC) layers. Additionally, the architecture incorporates one input layer, eight leaky ReLU (LR) layers, four ReLU layers, five max pooling layers (MPL), six batch normalization (BN) layers, one cross‐channel normalization (CLN) layer, three dropout layers (DL), one softmax layer (SL), and one classification layer (CL). To enhance transparency and interpretability, the Local Interpretable Model‐Agnostic Explanations (LIME) method is integrated to describe the model's predictions. The DeepEBCDCNet model is assessed using two BC image datasets: BC (for detection) and Breast Cancer Histology Images (BACH) (for classification). A 10‐fold cross‐validation method ensures the reliability of the results. The model's performance is compared with state‐of‐the‐art hybrid approaches to assess its effectiveness in BC DAC. The model attained an accuracy of 98.53% in BC detection, while in BC classification (three‐class: Benign [B], In situ [IS], and Invasive Carcinoma [IC]), it achieved 98.33% accuracy. The proposed DeepEBCDCNet significantly reduces incorrect diagnoses and enhances classification accuracy, offering a reliable second opinion for pathologists in BC DAC.
We present a deep learning-based approach to match prediction in shogi, using convolutional neural networks to evaluate board states with high accuracy. A dataset of one million board states from professional matches is encoded into a structured binary representation, enabling efficient learning of position evaluations. After training, the model achieves an 88.6
Neurodegenerative diseases, such as Alzheimer and Parkinson, progressively impair cognitive abilities and motor skills, caused by the degradation of nerve cells. Cancer also remains a major global cause of mortality, driven by cellular mutation and degeneration. Since no definitive cures exist for these diseases, early diagnosis is critical to improving survival rates and enabling more effective treatments. Recent studies indicate that handwriting dynamics can serve as an informative biomarker for the early detection of neurodegenerative diseases. The Diagnosis AlzheimeR WIth haNdwriting (DARWIN) dataset includes handwriting samples from individuals diagnosed or not with Alzheimer, collected during predefined writing tasks. The Brain Tumor detection dataset comprises brain image scans of patients presenting three distinct cancer types. In this paper, we apply classification and explainability methodologies to both datasets to predict the presence of Alzheimer and to identify cases of brain cancer. Dimensionality reduction techniques are also employed to improve interpretability by concentrating the analysis on lower-dimensional feature spaces. Experimental results show that the time spent performing in-air handwriting movements constitutes the most influential feature for Alzheimer prediction. Additionally, for the cancer detection dataset, deep learning models combined with local explainability methods yield adequate results.
Structural crack detection plays a pivotal role in infrastructure health monitoring, yet faces challenges in handling complex textures and maintaining cross-domain generalization. This paper presents SAM-IND, a novel framework that synergizes the segment anything model (SAM) with implicit neural representations to address these challenges. Our approach introduces two key innovations: (1) A parameter-efficient adaptation strategy using low-rank adaptation (LoRA) matrices, enabling domain-specific feature learning while preserving SAM's generalization capabilities through frozen backbone parameters; (2) an implicit neural decoder that establishes continuous coordinate-to-segmentation mappings, effectively capturing high-frequency crack patterns. Extensive experiments on three benchmark datasets demonstrate SAM-IND's superior performance, achieving 89.63% F1 score on DeepCrack while requiring only 1.59M trainable parameters (1.75% of SAM's total). Notably, the model reduces cross-material generalization error by 8.74% compared to state-of-the-art methods, showing particular robustness in handling complex scenarios involving mottled surfaces and low-contrast cracks. This work provides a new paradigm for structural health monitoring that balances accuracy, efficiency, and adaptability in real-world inspection scenarios.
