3020 papers
This paper demonstrates that natural-language instructions are fragile under hidden operational requirements, revealing that while generic prompting fails to satisfy omitted constraints, performance improves significantly when those constraints are explicitly elicited and executed.
This paper presents a novel LLM-augmented decision intelligence framework that integrates FMEA-based risk quantification with large language model contextual reasoning to achieve statistically significant, high-recall real-time supply chain risk assessment, as validated by both parameterized simulations and extensive empirical testing on the DataCo Smart Supply Chain dataset.
This paper introduces a novel corpus of diplomatic counter-arguments from UN General Assembly speeches (2005–2024) and evaluates state-of-the-art large language models using temporally aligned fine-tuning strategies to establish a systematic benchmark for cross-domain generalization in counter-argument generation.
CUPID-Med is a novel framework that leverages conformal uncertainty-aware prototype discovery to generate clinically interpretable, set-valued cluster predictions with calibrated confidence scores, significantly outperforming deterministic baselines in structuring unlabeled medical image data.
This study proposes an explainable, two-layered game-theoretic deep learning framework that combines SHAP for feature interpretability and Nash equilibrium for competitive pricing analysis to significantly outperform traditional models in forecasting short-term rental demand.
This paper proposes a "Smart Learning Estimator" (SLE), a deep learning-augmented Kalman filter combining TGAN, LSTM, and RWFNN architectures, which significantly enhances the accuracy and robustness of low-cost INS/GNSS navigation systems during extended GNSS signal outages by overcoming the limitations of traditional linear filters.
This paper proposes TDF-LSTGBM, a novel two-stage hybrid architecture that extracts temporal features from the penultimate layer of a stacked LSTM and combines them with consumer and temporal attributes for a Gradient Boosting Machine, achieving significantly higher accuracy in multi-granularity smart grid electricity demand forecasting across five consumer categories compared to traditional LSTM and ARIMA models.
GraphGarment is a data-efficient sim-to-real framework that combines a simulation-trained graph neural network with a lightweight residual corrector to accurately predict garment dynamics from minimal real-world data, enabling robust robotic garment hanging.
This study presents a unified, explainable deep learning framework that integrates ResNet50-based microstructure classification and YOLOv8-based surface defect detection with rule-based severity estimation to enable robust, interpretable, and automated quality control for steel manufacturing.
This paper proposes a yarn-level simulation method that leverages woven patterns to compute directional heat diffusion and shrinkage gradients, thereby accurately generating complex thermal buckling, twisting, and nonuniform wrinkle behaviors in fabrics that traditional surface-level models fail to capture.