927 papers
This paper demonstrates that a decision-aware machine learning framework for academic early warning outperforms traditional rule-based and probability-threshold approaches by reducing missed interventions and narrowing equity disparities through context-sensitive support decisions derived from calibrated risk estimates.
This paper proposes an AI-driven framework that integrates predictive maintenance and process optimization to enhance operational efficiency and system reliability in industrial manufacturing by leveraging real-time sensor data for fault detection, Remaining Useful Life estimation, and dynamic parameter adjustment.
This paper presents a machine learning study that successfully predicts professional volleyball match outcomes with over 93% accuracy using Gradient Boosting and Random Forest models trained on synthetic data and validated on real Polish PlusLiga matches, identifying OPI and Break-Point Differentials as the most critical predictive features.
This paper proposes an AI-driven trust-aware security framework for Cognitive Radio Networks that integrates Random Forest-based detection and Bayesian trust evaluation to effectively mitigate SSDF, PUE, and jamming attacks, achieving high detection accuracy and significantly improving network reliability and throughput even under severe attack conditions and low SNR levels.
This paper proposes CATBridge, a multi-stage knowledge enhancement framework that leverages a teacher-student collaboration mechanism to guide lightweight LLMs in achieving superior CVE-ATT&CK alignment accuracy in low-resource environments, significantly outperforming both state-of-the-art methods and zero-shot large-scale models.
This study evaluates hybrid classical-quantum routing for a 10-node space-based data center using a CUDA-Q workflow, finding that while photonic backends achieve superior accuracy with low latency, variational quantum circuit (VQC) backends incur prohibitive latency penalties despite offering richer quantum-state structures.