863 papers
This scoping review analyzes existing IoT security frameworks for digital health systems, revealing that while current solutions often address discrete areas, they lack comprehensive integration, and it identifies key elements like vulnerability assessments, threat modelling, and AI as essential for developing robust, proactive defense mechanisms.
This paper presents a systematic ablation study on adapting TrOCR for medieval handwritten text recognition, demonstrating that specific layer freezing strategies and the removal of contrast normalization can achieve competitive accuracy on small historical datasets while providing cross-dataset validation and interpretability insights.
This study proposes a deep learning framework using the UNSW-NB15 dataset for network intrusion detection and multi-class attack classification, demonstrating that an RNN-LSTM architecture outperforms ANN and CNN models by achieving a 97.42% accuracy through optimized data preprocessing and class balancing techniques.
This paper demonstrates that interpretable machine learning models trained on public ClinicalTrials.gov protocol features can accurately predict the probability of technical and regulatory success for small-molecule trials, outperforming standard benchmarks and offering a reproducible tool for risk-adjusted valuation.
The paper proposes SupCon-Mamba, a few-shot hyperspectral target detection framework that integrates a local context encoding mechanism, a pyramid Mamba module for efficient multi-scale spectral modeling, and supervised contrastive learning to eliminate false negatives and achieve superior detection performance with minimal labeled samples.
LGAgent is a structured, rule-guided multi-agent framework that leverages collaborative interactions between LLMs of varying scales to mimic professional legal workflows, thereby delivering interpretable, verifiable, and superior adjudicative reasoning compared to standard single-model approaches.
This paper introduces AIMAPP, a self-supervised, ROS-compatible framework that unifies online mapping, localisation, and planning for autonomous robots in unfamiliar environments by leveraging Active Inference to dynamically build sparse topological maps and balance exploration with goal-directed navigation without relying on predefined maps or pre-training.
This paper proposes a hybrid Bitcoin forecasting engine that integrates Permutation Entropy-guided dual-layer signal decomposition (CEEMDAN-VMD) with a Multi-Swarm Optimization-optimized Deep Neural Network, demonstrating statistically significant superior performance in accuracy and directional prediction compared to baseline models and single-swarm alternatives.
This paper proposes a contribution-aware federated edge learning algorithm (CA-FE-MADDPG) that integrates spatial interference modeling, fine-grained penalty mechanisms, and heuristic-guided initialization to address spectrum interference, unfair credit allocation, and environmental non-stationarity in massive IoT networks, thereby significantly improving system throughput, access rates, and quality of service.
This paper proposes ADEST-U-Net, a dual-encoder architecture that leverages cross-organ and cross-modality transfer learning from MRI brain tumor models to significantly improve liver tumor segmentation accuracy and reduce false positives in CT scans under data-scarce conditions.