927 papers
MobiShield is a novel federated learning framework for Android malware detection that simultaneously addresses non-IID data heterogeneity, concept drift, and privacy leakage by integrating adaptive feature weighting, Hoeffding-based drift detection, and a dynamic differential privacy mechanism that adjusts noise levels in response to distribution shifts.
This paper proposes a privacy-preserving federated learning framework that integrates a hybrid Deep Neural Network and XGBoost model to effectively detect cyber threats in Internet of Medical Things environments while addressing challenges related to data privacy, non-IID data distribution, and adversarial attacks.
This paper proposes a lightweight, modular retrieval-grounded framework that effectively detects and corrects hallucinations in Large Language Models by leveraging TF-IDF evidence retrieval and weighted verification, achieving high detection accuracy and significant hallucination reduction without the computational overhead of multi-LLM inference.
This paper proposes a novel conceptual framework, Secure QoS-IPv6, that synergistically integrates AI-driven anomaly detection, a lightweight consortium blockchain, and Deep Reinforcement Learning-based routing within a four-layer SDN and 6TiSCH architecture to holistically address security, trust, and Quality of Service challenges in resource-constrained, heterogeneous IPv6 IoT networks.
This study demonstrates that citation-risk outcomes for clinical gastroenterology articles can be accurately predicted using only publication-time information, with nonsemantic baselines and whole-document embeddings achieving high performance in identifying low-citation papers within a bounded literature corpus.
The paper proposes CAFR, a two-path detection framework that enhances unknown object detection by employing an asymmetric rectification mechanism to prevent known-category overconfidence and prototype-guided foreground modeling to improve objectness estimation, thereby achieving superior precision and F1 scores while maintaining stable known-category performance.
This paper proposes a learning-driven autonomy framework for safety-critical open-architecture autopilots that integrates a triad-conditioned CVAE for uncertainty-aware forecasting with risk-aware MPC and Control Barrier Function safety filtering, all orchestrated within a heterogeneous triple-redundant architecture to ensure safe, adaptive vehicle navigation.
This paper proposes a dual-channel patent vectorization method that leverages claim structure analysis to separately encode system-level and detail-level semantics, which are then fused via a gated mechanism with decorrelation correction to significantly improve patent matching accuracy over existing full-text baselines.
This paper presents MotionMonitor, a technology probe combining smartwatch sensing and mobile interfaces to support clinicians and educators in interpreting ADHD-related behaviors within classroom contexts, emphasizing that effective wearable systems must preserve environmental nuance, minimize student distraction, and serve as prompts for human sensemaking rather than automated diagnostic tools.
This paper proposes a domain-specific diffusion-based data augmentation framework that generates annotation-consistent synthetic images for rare pest species on sticky traps, effectively bridging the data scarcity gap and significantly improving instance segmentation performance across multiple architectures without requiring additional manual labeling.