2890 papers
This paper introduces an adaptive interpolatory subdivision scheme that learns local insertion angles via a compact edge-wise predictor to outperform fixed-tension baselines in accuracy and smoothness across Euclidean, spherical, and hyperbolic geometries while preserving the guarantee that refined curves pass through original control points.
This paper presents X-WAF, a hybrid machine learning framework that combines an ensemble of classifiers with SHAP-driven explainability to achieve high-accuracy, low-latency detection of zero-day web attacks while automatically translating insights into actionable ModSecurity rules for seamless integration into existing security infrastructures.
This paper introduces StrokeID-NER, a publicly available Indonesian dataset of 1,742 patient-generated health queries annotated for stroke-related named entities, which was used to demonstrate that fine-tuned transformer models significantly outperform zero-shot baselines in recognizing informal and regional medical terminology, while highlighting that future performance gains depend primarily on expanding lexical coverage rather than model architecture.
This study introduces a scalable, theory-grounded Agentic AI framework that rigorously evaluates the credibility of corporate SDG claims against a five-dimensional rubric, revealing that while most disclosures are partially substantive, firms often under-claim their actual sustainability engagement.
This study proposes an improved Mask R-CNN framework enhanced by histogram equalization and the Adam optimizer to achieve a 90.4% mAP in road pothole detection and instance segmentation, significantly outperforming the baseline model for intelligent road monitoring.
This paper demonstrates that class-conditional denoising diffusion probabilistic models (DDPMs) effectively mitigate severe class imbalance in predictive maintenance by synthesizing high-fidelity minority-class samples, significantly outperforming traditional augmentation methods like SMOTE and GANs in detecting rare industrial failures.
This paper introduces Causal Retrieval-Augmented Generation (CR-RAG), a framework that models the RAG pipeline as a Structural Causal Model to mitigate spurious correlations and reduce hallucinations through backdoor adjustment, inverse-propensity reweighting, and counterfactual data augmentation, achieving significant performance gains on knowledge-intensive benchmarks, particularly for long-tail questions.
This paper proposes a novel diffusion framework for image restoration that employs interactive modeling of determinism and stochasticity to effectively adapt to the heterogeneous requirements of diverse degradation tasks, achieving superior performance and interpretability across multiple restoration scenarios.
This paper proposes a topology-preserving, direction-aware adaptive framework for yarn-level cloth simulation that dynamically adjusts constraint computation intensity based on local deformation and contact activity while employing visual upsampling to maintain high-fidelity woven structures without incurring prohibitive computational costs.
This paper proposes IBKA-Transformer, a novel fault diagnosis method that integrates an improved Black Kite Algorithm with a Transformer model to optimize hyperparameters and select features, achieving 99.86% accuracy in industrial robot bearing fault diagnosis.