2608 papers
This paper proposes a Weighted Sparse Matrix Regression (WSMR) model that integrates structural and distance information to effectively handle continuous occlusion in face recognition, utilizing the Alternating Direction Method of Multipliers (ADMM) for optimization and demonstrating superior performance through experiments on public databases.
This paper proposes a CKKS-based homomorphic federated learning framework that enables privacy-preserving disease prediction across multiple hospitals by encrypting model updates to prevent membership inference attacks, though this enhanced security comes at the cost of significant communication overhead and reduced predictive utility compared to plaintext approaches.
This paper proposes and validates a label-free geometric diagnostic that uses the angle between domain subspaces and class-discriminative directions to predict when removing domain-specific information will improve out-of-distribution seed classification, particularly under severe deployment shifts where traditional selection methods fail.
This study found that the EduQubit learning environment, which structures exploration through guided missions, significantly improved conceptual understanding of quantum computing among sixth-grade students compared to the IBM Quantum Composer, though it did not yield significant differences in self-efficacy, workload, or user experience.
This paper proposes an AI-enabled framework that integrates machine learning classification with ADMM-LASSO optimization to enhance the resilience, workload balancing, and resource efficiency of enterprise instant messaging systems under dynamic conditions.
This paper proposes a closed-loop V2X security and energy allocation framework that integrates an MTCNN-LSTM-based multilevel trust evaluation model to achieve high-accuracy multi-attack detection and optimize battery consumption in smart grid-enabled electric vehicle networks.
This paper proposes ACO-QSNR, a novel cross-layer routing protocol for Mobile Ad Hoc Networks that integrates Ant Colony Optimization and Q-learning with SNR-aware metrics to outperform conventional protocols like AODV in terms of packet delivery ratio, throughput, and delay under dynamic mobility and varying node densities.
This paper proposes a geometric framework for operationalizing conceptual scope in scientific texts using transformer embeddings, demonstrating that geometric deviation from topic-specific contexts correlates with language-model surprisal and is consistently captured even in compact abstracts.
EcoAccel-ITAD is a rapid, telemetry-driven framework that leverages vendor-specific APIs to assess the health and remaining service life of decommissioned AI accelerators, thereby enabling precise quantification of avoided embodied carbon for their second-life deployment.
This study develops and deploys an interactive Streamlit dashboard featuring a Random Forest classifier enhanced with engineered risk features to accurately predict bank customer churn, achieving high performance metrics and providing actionable, explainable retention insights for relationship managers.