2661 papers
This study presents a fully deployed, interactive Streamlit dashboard that utilizes an XGBoost classifier trained on 180,519 DataCo Global logistics records to predict late delivery risks with 79.89% accuracy, enabling proactive operational decisions in shipping, regional monitoring, and customer communication.
This paper presents xLaDe, a Lean 4 ecosystem framework designed to ensure the long-term reproducibility and sustainability of formalized mathematics by recording toolchain metadata and managing experimental environments without modifying source code, thereby addressing the challenges posed by the language's rapid evolution.
This paper demonstrates that replacing verbalized confidence with a budget-matched token-probability abstention policy causally eliminates a specific accuracy-degrading effect observed in the llama3.2:1b model, but fails to generalize this benefit to other small open-weight models, indicating that such fixes are model-specific and that fixed-budget abstention can remain net-harmful even with improved targeting.
This study proposes an explainable Digital Twin framework that integrates machine learning, counterfactual simulation, and cost-aware optimization to model multidimensional poverty dynamics in Nigeria's semi-urban Isolu community, demonstrating that targeted, cluster-specific multi-policy interventions are more effective than uniform safety nets for intelligent poverty governance.
This paper presents a multi-modal testing framework that decouples perception from control by converting photorealistic inputs into abstract line drawings and semantic maps, validated through a novel open simulator that bridges fast training-level fidelity with rigorous register-level firmware emulation to expose defects invisible to standard simulation.
This paper proposes an Adaptive Hybrid Collaborative Filtering (AHCF) model that integrates offline matrix factorization and topic modeling with an online adaptive clustering mechanism to significantly reduce retraining time while maintaining competitive prediction accuracy for real-time, large-scale recommendation systems.
This paper introduces an auditable three-gate evaluation framework demonstrating that while fallible verifiers can recover from model stopping errors via resampling, current methods fail to identify the optimal action selection between resampling and rerouting, thereby establishing bounded recovery potential without supporting a complete policy-learning chain.
This study addresses the gap in applying reinforcement learning to OSAT manufacturing by developing an executable benchmark and decision framework that reveals adaptive scheduling policies offer selective, objective-specific advantages over traditional heuristics rather than a universal replacement, thereby guiding managers on the trade-offs and requirements for operational deployment in Vietnam.
This paper proposes a deep-learning-based method for identifying transformed immersive video content by utilizing viewport-aware preprocessing and transformation-robust feature points, achieving a 97.5% recognition rate and improved computational efficiency without relying on metadata.
This study evaluates the transformation-specific robustness of machine learning-based network intrusion detection systems on the UNSW-NB15 dataset, revealing that while models exhibit varying sensitivity to feature perturbations, a TTL-aware training configuration significantly reduces performance degradation under negative time-to-live modifications compared to a standard baseline.