2786 papers
This paper proposes a hybrid t-SNE-UMAP optimization method that leverages UMAP's graph structure to initialize and reinforce t-SNE's local neighborhood preservation, resulting in statistically significant improvements in trustworthiness and neighborhood recall across nine datasets despite a trade-off in density preservation.
This paper proposes a training-free framework that fuses dense monocular depth priors with sparse, high-confidence stereo geometric anchors via cross-modal scale alignment and edge-aware propagation to achieve robust, zero-shot depth perception that outperforms existing methods in edge-region accuracy.
The paper proposes ADF-Pose, an agent-driven framework featuring an Identity Prior Extractor and a Multimodal Identity Converter to effectively preserve facial identity and fine-grained details in pose-guided person image generation, particularly under large pose variations where existing methods struggle.
The paper proposes CDA-YOLO, an enhanced YOLOv11 framework incorporating a Cross-level Bidirectional Feature Pyramid Network, a lightweight Dual-Scale Convolution module, and a parameter-free average attention mechanism to significantly improve the accuracy of PCB defect detection by addressing challenges with tiny targets and complex backgrounds.
This study presents an explainable, leakage-aware framework for district-level wheat yield forecasting in Madhya Pradesh, India, demonstrating that rigorous validation protocols and the inclusion of historical yield data are more critical to predictive accuracy than model complexity, with Elastic Net achieving the best protected-test performance.
This study reveals that while current generative AI models demonstrate high internal stability, they exhibit lower agreement with human raters, systematic severity biases, and failure to apply specific rubric rules compared to human double-marking, suggesting they are best suited for supervised, discrepancy-triggered assistance rather than autonomous replacement in consequential EFL writing assessment.
This paper proposes an integrated hybrid framework that combines TF-IDF, semantic similarity, and collaborative filtering for candidate provider selection with a trust-aware review intelligence system using Multi-Task BiLSTM and Isolation Forest to analyze aspect-specific sentiment and review credibility, thereby achieving more reliable service provider matching through a unified re-ranking pipeline.
This paper demonstrates that a compact 201-million-parameter language model using deterministic, token-identity-based multi-hash routing for expert selection can achieve competitive performance on long-horizon training tasks, though the specific contribution of the routing mechanism remains unisolated due to the lack of matched dense controls.
This paper introduces Fed-DLC, a lightweight federated learning framework that employs a dynamic cipher to secure model exchanges in resource-constrained networks, demonstrating that it maintains task performance comparable to standard federated learning while adding manageable computational overhead, though its cryptographic security requires further validation.
This study introduces a reproducible, bilingual NLP pipeline that adapts the Say-Do Gap Index architecture to detect divergence between corporate rhetoric and observable governance practices regarding agentic AI within the IBEX 35 index, demonstrating the framework's generalizability beyond traditional ESG domains.