2931 papers
This paper proposes MEFP-Net, a multimodal recognition algorithm utilizing hierarchical mid-stage fusion and a dual-path backbone to effectively address feature imbalance and lighting sensitivity, thereby significantly improving the detection of small and occluded objects in complex urban traffic scenarios.
This paper presents a comprehensive experimental evaluation of various Facebook AI Similarity Search (FAISS) indexing strategies, analyzing their trade-offs in accuracy, latency, and memory usage across different distance metrics and quantization techniques to provide practical guidance for large-scale similarity search deployments.
This paper proposes and evaluates a predictive, adaptive scheduling model that integrates workload forecasting with Ceph-aware storage decisions to significantly reduce resource contention and I/O latency in Kubernetes–Ceph hyperconverged infrastructure running on Proxmox VE, achieving superior load balancing and performance compared to the default Kubernetes scheduler.
This paper proposes a federated reinforcement learning framework for mobile edge networks that combines a VAE-LSTM model for content popularity prediction with a multi-agent deep reinforcement learning algorithm to optimize cooperative caching decisions, thereby significantly reducing latency and improving cache hit rates compared to existing baseline methods.
This paper proposes DMPB-Net, a novel semi-supervised framework that integrates a Mamba-driven encoder with specialized modules for semantic decoupling, prototype attention, and boundary refinement to achieve state-of-the-art breast lesion segmentation in ultrasound images under limited supervision.
This paper proposes a robust tri-modal framework that integrates CNNs, Vision Transformers, and Graph Neural Networks via an adaptive gated fusion mechanism to achieve state-of-the-art forgery detection accuracy (99.10%) and interpretability on social media images.
This study proposes a principled, three-part decision framework for selecting models in data-limited scientific machine learning, demonstrating through kinase inhibitor and solubility case studies that simple linear models often outperform complex alternatives when sample sizes are small, thereby advocating for model simplicity as a critical baseline.
This study evaluates various machine learning regression models on the Individual Household Electric Power Consumption dataset to predict energy trends and identify anomalies, finding that Random Forest outperforms other algorithms while SHAP analysis reveals Global_intensity as the dominant predictive factor.
This paper proposes the Stigmergic Skeleton Field (SSF) framework, which integrates a medial-axis skeleton graph with an ant-colony pheromone field and a Localized Incremental Re-skeletonization (LIR) algorithm to enable efficient, decentralized multi-robot navigation in dynamic environments by repairing topology only where necessary, achieving near-optimal path quality with significantly faster re-planning speeds compared to existing baselines.
This paper presents Semperf, an LLM-assisted framework that constructs rank-profile matrices and clusters processes to enable scalable, automated performance diagnosis and bottleneck identification for extreme-scale parallel HPC applications.