2731 papers
This study validates the Explainable Multi-Stakeholder Educational Intelligence Framework (EMSEIF), an empirical system that transforms diverse higher education feedback into auditable, aspect-level evidence and actionable curriculum improvements through explainable AI, knowledge graphs, and fairness auditing.
This paper demonstrates that in automated paper self-review systems, predicting whether near-threshold decisions will flip under perturbation is effectively achieved by a simple baseline leveraging empirical score-shift distributions and decision margins, rather than by complex learned boundary models.
This paper proposes GlaucoFusionNet, a multimodal explainable AI framework that integrates retinal fundus images and clinical data using causal inference and counterfactual reasoning to overcome the limitations of correlation-driven models in providing interpretable, actionable glaucoma risk predictions and progression stratification.
This paper proposes a hybrid framework that combines Runge–Kutta optimized Neural Radiance Fields (NeRF) for geometry-aware 3D reconstruction with a diffusion-based refinement module to effectively correct severe selfie perspective distortion while preserving facial identity, structural consistency, and photorealistic quality.
This study proposes a lightweight, non-destructive ensemble CNN framework that achieves 97.50% accuracy in detecting eggshell cracks using multi-scale feature extraction, offering a computationally efficient and robust alternative to manual inspection and existing deep learning models for automated food packaging applications.
This scoping review maps the emerging landscape of "vibe coding" (LLM-assisted code generation) in healthcare across clinical, administrative, and research domains, highlighting its potential to democratize software development while identifying critical challenges in security and regulation and proposing a 3-tier governance framework for responsible adoption.
This paper presents HT-CS, a cost-sensitive incremental learning framework combining a Hoeffding Tree regressor with ADWIN drift detection and sample-level weighting, which significantly improves critical temperature event recall and prediction accuracy on resource-constrained Raspberry Pi 4B devices while maintaining low latency and a small model footprint.
This study proposes an improved AEP2-YOLOv8s model that integrates ROI extraction, background normalization, an AIFI module, EMA attention, and a four-scale detection architecture to achieve high-precision, non-destructive identification of micro-contaminants on eggshell surfaces, significantly outperforming the original YOLOv8s in accuracy.
The paper proposes DeepHistoViT, an interpretable Vision Transformer framework that achieves state-of-the-art accuracy and statistical robustness in classifying histopathological images for lung cancer, colon cancer, and acute lymphoblastic leukaemia by leveraging attention mechanisms to localize diagnostically relevant regions.
This paper proposes a Process-Oriented Fusion Network (PFN) that effectively addresses the challenges of multi-material, multi-stage yarn quality prediction by jointly modeling variable-length material compositions, nominal blending-ratio priors, and stage-specific process conditions, achieving superior performance on key quality indicators compared to existing methods.