2833 papers
This paper introduces the FC-CML algorithm, which leverages circulant topology and fast convolution to reduce the time and space complexity of chaotic neural image encoding from O(N³) and O(N²) to O(N log N) and O(N), respectively, thereby enabling stable, high-speed image processing on low-cost edge devices like the ESP32.
This paper proposes a Proximal Policy Optimization-based Reinforcement Learning framework that manages bidirectional Vehicle-to-Grid power flow using the Indian Grid Master dataset to balance economic and environmental benefits with battery longevity through an asymmetric reward function and strict State of Charge constraints.
This paper proposes an enhanced EfficientNetV2-L framework that utilizes a Stable Diffusion-generated balanced dataset to achieve superior accuracy and robustness in detecting deepfake lung CT images compared to existing deep learning models.
The paper introduces BioGCN, a biologically inspired Graph Convolutional Network that integrates ATC codes as domain knowledge to significantly improve the accuracy of predicting human oral bioavailability compared to traditional machine learning methods and existing models.
This paper proposes DTRW-YOLO11s, an enhanced insulator defect detection algorithm that integrates a DSDown downsampling module, Triplet Attention, and RFCAConv-based multi-scale perception with WIoU loss to significantly improve detection accuracy and computational efficiency while reducing false positives and missed detections in complex transmission line environments.
This paper proposes RECO-Rank, a compensatory multi-evidence ranking model that integrates semantic, topic, and citation signals to overcome the limitations of semantic-first filtering in cross-domain reviewer recommendation, thereby improving the identification of suitable reviewers with divergent terminology while maintaining top-ranked precision.
This paper proposes a novel, integrated AI framework combining neural networks, genetic algorithms, reinforcement learning, and support vector machines to simultaneously optimize cement mix design, kiln control, and lifecycle assessment, achieving significant reductions in CO₂ emissions, energy consumption, and raw material use while addressing key implementation challenges.
This paper proposes HCL-YOLOv8, a lightweight detection framework that enhances transmission-line insulator defect identification by integrating hierarchical consistency learning, task-aligned label assignment, and specialized loss functions to address challenges in localizing tiny defects within complex backgrounds.
This paper proposes a lightweight, interpretable framework that combines biomechanically guided standard motion modeling with a four-layer ST-GCN to provide phase-specific, high-accuracy quality assessment and feedback for the standing long jump in physical education settings.
This paper proposes a zero-shot adaptation framework for composed image retrieval that leverages structured semantic supervision from unlabeled images to improve fine-grained retrieval performance in annotation-scarce domains without relying on manually annotated triplets.