The increasing application of Artificial Intelligence (AI) in resource-limited environments, particularly within computer vision tasks, together with ethical and practical considerations, has raised the demand for lightweight models that are both efficient and interpretable. This survey explores the integration of lightweight neural networks with explainable AI (XAI), addressing the need for models that are both computationally efficient and transparent. It reviews lightweight neural network techniques along with their advantages and limitations, as well as widely used XAI methods. A Scopus-based analysis is conducted to identify the most widely adopted lightweight approaches in current literature. Among these techniques, knowledge distillation (KD) stands out as the most widely adopted; therefore, the survey examines how KD interacts with XAI. It also reviews a range of existing qualitative and quantitative evaluation metrics specifically designed for evaluating XAI effectiveness. Finally, it presents an illustrative comparison of explanations across Convolutional Neural Networks (CNNs) and Transformer architectures, demonstrating architectural differences that motivate the need for XAI-aware cross-architecture distillation methods. Despite advancements, challenges remain in fully eliminating trade-offs between model complexity and explainability. This survey highlights the need for improved evaluation metrics and the design of explainable lightweight models to support interpretable and deployable lightweight AI systems.

Federated learning (FL) has emerged as a privacy-preserving paradigm for collaborative intrusion detection in networked environments. However, it remains vulnerable to Poisoning Attacks (PA) wherein malicious clients can corrupt the global model through deceptive updates. To address this, we propose PoisonShield-FL-NIDS, a robust FL-based intrusion detection system that integrates client-side anomaly filtering with trust-aware aggregation to defend against poisoned contributions. Experimental evaluation under varying levels of adversarial influence demonstrates that PoisonShield-FL-NIDS achieves superior performance across key metrics, attaining 93% accuracy, 91% precision, 94% recall, and an AUC of 0.96, while maintaining a low robustness index RI < 0.05 even with 30% compromised clients. Compared to baseline FL models such as FL-CNN and FedACNN, our framework demonstrates faster convergence and higher resilience with only a marginal increase in communication overhead.

In this work, we present an innovative approach utilizing ControlNet-based diffusion models along with upscaling capabilities for domain adaptation and quality refinement of 3D modelled synthetic datasets, focusing on autonomous vehicle applications. A significant domain gap often exists between synthetic and real-world data, hindering the applicability of deep learning models trained on synthetic data for real-world scenarios. Our methodology leverages the strengths of Controlled Augmentation by simultaneously utilizing multiple ControlNet signals, including edge detection, depth information, segmentation maps, and tile resampling. To improve how synthetic data aligns with the desired domain specifications, these signals guide the generative process, and we also incorporate text-guided prompts extracted via Large Language Models (LLMs), to improve control over the synthesis of desired features and attributes. We test the approach on diverse environmental conditions from the VKITTI dataset, a well-known 3D modelled synthetic dataset generated in Unity for autonomous driving research. The refined data is validated using quantitative metrics including FID, SSIM, and LPIPS, and is also evaluated on downstream machine learning tasks of object detection and classification, using YOLO-v8 to ensure its utility and effectiveness. Experimental analysis demonstrates the effectiveness of this method in improving the realism and usability of synthetic data. Our approach contributes to fields that require high-quality data synthesis and domain adaptation. The experimental work, along with ControlNet models used in this project is available online.1

BACKGROUND: Recent developments in the domain of biomedical knowledge bases (KBs) open up new ways to exploit biomedical knowledge that is available in the form of KBs. Significant work has been done in the direction of biomedical KB creation and KB completion, specifically, those having gene-disease associations and other related entities. However, the use of such biomedical KBs in combination with patients' temporal clinical data still largely remains unexplored, but has the potential to immensely benefit medical diagnostic decision support systems. RESULTS: We propose two new algorithms, LOADDx and SCADDx, to combine a patient's gene expression data with gene-disease association and other related information available in the form of a KB, to assist personalized disease diagnosis. We have tested both of the algorithms on two KBs and on four real-world gene expression datasets of respiratory viral infection caused by Influenza-like viruses of 19 subtypes. We also compare the performance of proposed algorithms with that of five existing state-of-the-art machine learning algorithms (k-NN, Random Forest, XGBoost, Linear SVM, and SVM with RBF Kernel) using two validation approaches: LOOCV and a single internal validation set. Both SCADDx and LOADDx outperform the existing algorithms when evaluated with both validation approaches. SCADDx is able to detect infections with up to 100% accuracy in the cases of Datasets 2 and 3. Overall, SCADDx and LOADDx are able to detect an infection within 72 h of infection with 91.38% and 92.66% average accuracy respectively considering all four datasets, whereas XGBoost, which performed best among the existing machine learning algorithms, can detect the infection with only 86.43% accuracy on an average. CONCLUSIONS: We demonstrate how our novel idea of using the most and least differentially expressed genes in combination with a KB can enable identification of the diseases that a patient is most likely to have at a particular time, from a KB with thousands of diseases. Moreover, the proposed algorithms can provide a short ranked list of the most likely diseases for each patient along with their most affected genes, and other entities linked with them in the KB, which can support health care professionals in their decision-making.

Abstract A graph can represent a complex organization of data in which dependencies exist between multiple entities or activities. Such complex structures create challenges for machine learning algorithms, particularly when combined with the high dimensionality of data in current applications. Graph convolutional networks were introduced to adopt concepts from deep convolutional networks (i.e. the convolutional operations/layers) that have shown good results. In this context, we propose two major enhancements to two of the existing graph convolutional network frameworks: (1) topological information enrichment through clustering coefficients; and (2) structural redesign of the network through the addition of dense layers. Furthermore, we propose minor enhancements using convex combinations of activation functions and hyper-parameter optimization. We present extensive results on four state-of-art benchmark datasets. We show that our approach achieves competitive results for three of the datasets and state-of-the-art results for the fourth dataset while having lower computational costs compared to competing methods.

One of the key challenges in systems biology is to derive gene regulatory networks (GRNs) from complex high-dimensional sparse data. Bayesian networks (BNs) and dynamic Bayesian networks (DBNs) have been widely applied to infer GRNs from gene expression data. GRNs are typically sparse but traditional approaches of BN structure learning to elucidate GRNs often produce many spurious (false positive) edges. We present two new BN scoring functions, which are extensions to the Bayesian Information Criterion (BIC) score, with additional penalty terms and use them in conjunction with DBN structure search methods to find a graph structure that maximises the proposed scores. Our BN scoring functions offer better solutions for inferring networks with fewer spurious edges compared to the BIC score. The proposed methods are evaluated extensively on auto regressive and DREAM4 benchmarks. We found that they significantly improve the precision of the learned graphs, relative to the BIC score. The proposed methods are also evaluated on three real time series gene expression datasets. The results demonstrate that our algorithms are able to learn sparse graphs from high-dimensional time series data. The implementation of these algorithms is open source and is available in form of an R package on GitHub at https://github.com/HamdaBinteAjmal/DBN4GRN, along with the documentation and tutorials.

In a Convolutional Neural Network, each neuron in the output feature map takes input from the neurons in its receptive field. This receptive field concept plays a vital role in today's deep neural networks. However, inspired by neuro-biological research, it has been proposed to add inhibitory neurons outside the receptive field, which may enhance the performance of neural network models. In this paper, we begin with deep network architectures such as VGG and ResNet, and propose an approach to add lateral inhibition in each output neuron to reduce its impact on its neighbours, both in fine-tuning pre-trained models and training from scratch. Our experiments show that notable improvements upon prior baseline deep models can be achieved. A key feature of our approach is that it is easy to add to baseline models; it can be adopted in any model containing convolution layers, and we demonstrate its value in applications including object recognition and semantic segmentation of aerial images, where we show state-of-the-art result on the Aeroscape dataset. On semantic segmentation tasks, our enhancement shows 17.43% higher mIoU than a single baseline model on a single source (the Aeroscape dataset), 13.43% higher performance than an ensemble model on the same single source, and 7.03% higher than an ensemble model on multiple sources (segmentation datasets). Our experiments illustrate the potential impact of using inhibitory neurons in deep learning models, and they also show better results than the baseline models that have standard convolutional layer.

Accidents and attacks that involve chemical, biological, radiological/nuclear or explosive (CBRNE) substances are rare, but can be of high consequence. Since the investigation of such events is not anybody's routine work, a range of AI techniques can reduce investigators' cognitive load and support decision-making, including: planning the assessment of the scene; ongoing evaluation and updating of risks; control of autonomous vehicles for collecting images and sensor data; reviewing images/videos for items of interest; identification of anomalies; and retrieval of relevant documentation. Because of the rare and high-risk nature of these events, realistic simulations can support the development and evaluation of AI-based tools. We have developed realistic models of CBRNE scenarios and implemented an initial set of tools.

Many businesses rely on forecasting techniques to detect whether sales opportunities are likely to be won or at risk of being lost. This enables the businesses to respond proactively. This paper describes a new method of sales forecasting that improves on an existing Qualitative Sales Predictor (QSP) in Hewlett-Packard (HP). QSP is based on a series of qualitative assessments that are made by sales personnel, the results of which are combined using weighted factors. In this research, we have developed an alternative method of forecasting sales opportunities, with three key differences: (1) the qualitative assessments are supplemented with quantitative data describing attributes of the opportunity, (2) we replace the weight factors with a Tree Augmented Naïve Bayes (TAN) classifier that can capture dependences between variables and produces a probabilistic output to which thresholds can be applied, (3) the TAN classifier is of course learned from historical data, whereas the existing QSP has fixed weights. Our approach has an accuracy of 90.6% in predicting whether sales will be won or lost, a substantial improvement on the existing approach's accuracy of 75.6% on the same unseen test data.

Similarity, and the strongly related inverse concept of distance, plays an important role in many data mining techniques. Typically, the definitions of similarity/distance measures are restricted to being symmetric. In this paper, however, we consider asymmetric measures and show that there are justifiable reasons in many domains for preferring asymmetric measures. Our review of the literature indicates that these ideas have occasionally been considered in various guises, but not formalized. In this context, we use our proposed Contains concept to represent an asymmetric relation. We show how asymmetry can be introduced into some widely used binary similarity/distance measures and evaluate their performance on data sets from multiple domains. The results show that asymmetric Contains measures can yield better performance, and are never worse than the corresponding symmetric measures.

In current state-of-the-art commercial first person shooter games, computer controlled bots, also known as nonplayer characters, can often be easily distinguishable from those controlled by humans. Tell-tale signs such as failed navigation, “sixth sense” knowledge of human players' whereabouts and deterministic, scripted behaviors are some of the causes of this. We propose, however, that one of the biggest indicators of nonhumanlike behavior in these games can be found in the weapon shooting capability of the bot. Consistently perfect accuracy and “locking on” to opponents in their visual field from any distance are indicative capabilities of bots that are not found in human players. Traditionally, the bot is handicapped in some way with either a timed reaction delay or a random perturbation to its aim, which doesn't adapt or improve its technique over time. We hypothesize that enabling the bot to learn the skill of shooting through trial and error, in the same way a human player learns, will lead to greater variation in game-play and produce less predictable nonplayer characters. This paper describes a reinforcement learning shooting mechanism for adapting shooting over time based on a dynamic reward signal from the amount of damage caused to opponents.

Automated screening systems are commonly used to detect some agent in a sample and take a global decision about the subject (e.g., ill/healthy) based on these detections. We propose a Bayesian methodology for taking decisions in (sequential) screening systems that considers the false alarm rate of the detector. Our approach assesses the quality of its decisions and provides lower bounds on the achievable performance of the screening system from the training data. In addition, we develop a complete screening system for sputum smears in tuberculosis diagnosis, and show, using a real-world database, the advantages of the proposed framework when compared to the commonly used count detections and threshold approach.

We describe an application of probabilistic modeling and inference technology to the problem of analyzing sensor data in the setting of an intensive care unit (ICU). In particular, we consider the arterial-line blood pressure sensor, which is subject to frequent data artifacts that cause false alarms in the ICU and make the raw data almost useless for automated decision making. The problem is complicated by the fact that the sensor data are acquired at fixed intervals whereas the events causing data artifacts may occur at any time and have durations that may be significantly shorter than the data collection interval. We show that careful modeling of the sensor, combined with a general technique for detecting sub-interval events and estimating their duration, enables effective detection of artifacts and accurate estimation of the underlying blood pressure values.

This paper outlines a system for detection of cardiac arrhythmias within ECG signals, based on a Bayesian artificial neural network (ANN) classifier. The Bayesian (or probabilistic) ANN classifier is built by the use of a logistic regression model and the backpropagation algorithm based on a Bayesian framework. Its performance for this task is evaluated by comparison with other classifiers including Naive Bayes, decision trees, logistic regression, and RBF networks. A paired t-test is employed in comparing classifiers to select the optimum model. The system is evaluated using noisy ECG data, to simulate a real-world environment. It is hoped that the system can be further developed and fine-tuned for practical application.

We consider a dynamic market-place of self-interested agents with differing capabilities. A task to be completed is proposed to the agent population. An agent attempts to form a coalition of agents to perform the task. Before proposing a coalition, the agent must determine the optimal set of agents with whom to enter into a coalition for this task; we refer to this activity as coalition calculation. To determine the optimal coalition, the agent must have a means of calculating the value of any given coalition. Multiple metrics (cost, time, quality etc.) determine the true value of a coalition. However, because of conflicting metrics, differing metric importance and the tendency of metric importance to vary over time, it is difficult to obtain a true valuation of a given coalition. Previous work has not addressed these issues. We present a solution based on the adaptation of a multi-objective optimization evolutionary algorithm. In order to obtain a true valuation of any coalition, we use the concept of Pareto dominance coupled with a distance weighting algorithm. We determine the Pareto optimal set of coalitions and then use an instance-based learning algorithm to select the optimal coalition. We show through empirical evaluation that the proposed technique is capable of eliciting metric importance and adapting to metric variation over time.