3020 papers
This paper presents a privacy-preserving, resource-constrained framework for early student risk prediction that integrates calibrated XGBoost models, SHAP-based explainability, and a budget-aware agentic intervention layer to ensure accurate, interpretable, and compliant decision-making while supporting machine unlearning.
This paper introduces CerebroTune, a novel bio-inspired hybrid metaheuristic that integrates Cellular Automata, a modified Artificial Bee Colony algorithm, and Reinforcement Learning with Hopfield networks to dynamically optimize deep learning hyperparameters, achieving superior accuracy, reduced computational costs, and enhanced stability across diverse domains compared to existing methods.
This paper proposes an Improved Gold Mine Optimization Algorithm with a dual-layer information-sharing framework and dynamic Pareto dominance (IGMO-DP) to effectively solve the multi-objective scheduling problem of heterogeneous UAV swarms by simultaneously optimizing mission completion time, energy consumption, and load balance, demonstrating superior convergence and solution diversity compared to state-of-the-art algorithms.
This paper argues that the traditional citation-based metric for scientific impact is becoming obsolete in an era of unbundled journals and AI-mediated knowledge access, necessitating a new system that recognizes and measures the contributions of AI engagement, open-source libraries, and datasets that lack formal publications.
This paper proposes a queue-aware task offloading method for the computing continuum that hybridizes an analytically corrected G/G/c delay approximation with real-time queue observations to dynamically assign tasks under variable workloads, thereby significantly reducing latency and estimation error compared to existing baselines without requiring training or calibration.
This paper proposes DPENet, a dynamic perception enhancement network that integrates adaptive dynamic convolution, deformable attention, and dynamic sampling modules to effectively address challenges like varying target sizes, dense small objects, and complex backgrounds in high-resolution remote sensing image detection.
This study introduces a novel MACB-ResNet deep learning model that significantly improves the accuracy of classifying virtual taste stimuli from facial surface electromyography (sEMG) signals to 86.32%, outperforming traditional and standard deep learning classifiers for applications in digital health and food technology.
This study proposes a transfer learning-based framework that enhances BiSeNetV2 with a hybrid Cross-Entropy–Lovasz loss to achieve robust, real-time seam segmentation for autonomous robotic welding in construction, significantly improving performance on reflective surfaces without increasing computational complexity.
TITAN is a novel three-tier federated learning architecture for Industrial IoT intrusion detection that overcomes non-IID data and class imbalance challenges by exchanging only prediction outputs and device summaries via adaptive knowledge distillation and Bayesian aggregation, achieving superior macro-F1 performance over existing methods without requiring raw data or gradient sharing.
This paper proposes a hierarchical federated learning framework for healthcare that integrates client-level differential privacy with institution-level CKKS homomorphic encryption, demonstrating that while differential privacy incurs measurable accuracy losses dependent on the privacy budget, homomorphic encryption introduces no utility degradation, thereby isolating and quantifying the distinct privacy-utility tradeoffs of each mechanism.