2731 papers
This paper demonstrates that while selecting an optimal composition of heterogeneous ensemble members yields marginal gains over a uniform mix, it fails to outperform simply choosing the best single family, revealing that the perceived potential for improvement is largely an artifact of selection bias rather than genuine untapped performance.
This study evaluates the impact of duplicate exposure and candidate-coverage stress testing on the Clinical Abbreviation Sense Inventory (CASI), revealing that while data leakage had a negligible effect on aggregate accuracy, high in-support confidence calibration failed to reliably detect cases where the correct sense was missing from the candidate inventory.
This paper introduces Agent Tree (A-Tree), an interpretable extension of Neural Network Trees that utilizes specialized agents and a one-step look-ahead mechanism to enhance decision transparency and accuracy in high-stakes domains, demonstrated through a successful application in classifying and mining innovation rules from computer-vision patents.
This paper introduces MAFL-PdM, a multi-agent federated learning framework that leverages post-federation personalisation to significantly outperform both centralised training and federated baselines in predicting equipment remaining useful life across heterogeneous industrial IoT sites while preserving data privacy.
This paper demonstrates that a fully unprivileged co-tenant on a shared cloud FPGA can recover a complete AES-128 master key from a victim by exploiting DDR4 row-buffer timing side-channels, overcoming the "one-slot-per-bank" limitation through reverse-engineered bank mapping and statistical aggregation across multiple encryptions.
This paper proposes a Hybrid Deep Ensemble Intrusion Detection System (HDE-IDS) that integrates a Transformer-enhanced DNN with XGBoost and adaptive Zero Trust security to significantly improve malware detection accuracy, stability, and real-time resilience for Industrial IoT infrastructure.
This paper proposes a semantic-enhanced multi-task framework with structure-aware curriculum learning, leveraging Qwen2-7B-Instruct and generative Semantic Role Labeling to significantly improve low-resource neural machine translation performance by resolving semantic misalignment and optimizing training dynamics.
This paper introduces the Dynamic Adaptive Multimodal Graph Fusion Network (DAMGF-Net), a novel deep learning model that dynamically integrates remote sensing, social media, IoT, and seismic data to achieve highly accurate, interpretable, and near-real-time multi-hazard recognition across earthquakes, floods, fires, and smoke.
This study introduces FZ-VLM, a novel two-stage Florence-Zephyr Vision-Language Model framework that automates the extraction of radiological attributes from lung CT scans and generates clinically relevant nodule descriptions and follow-up recommendations, demonstrating superior accuracy and completeness compared to existing AI baselines and human experts.
This paper introduces TurkCite, a framework and expert-annotated dataset for classifying citation intents in Turkish academic literature, demonstrating that language-specific Transformer encoders outperform translation-based transfer while offering prompt-optimized large language models as effective training-free alternatives for non-English bibliometric analysis.