2608 papers
This paper proposes SCGS, a lightweight and efficient 3D reconstruction method for power equipment that leverages semantic-aware contour guidance to overcome background clutter and material complexity using only smartphone images, achieving significantly faster processing and higher fidelity than existing NeRF-based approaches.
This paper exposes a fatal label-feature circularity in standard risk-based software testing that inflates machine learning performance, then proposes a rigorous protocol using leaky-feature removal and strict evaluation to demonstrate a modest but statistically robust 3.64% improvement over strong baselines while revealing that these models fail to generalize temporally.
This paper demonstrates that short, automatically optimized prefixes can significantly compromise the safety alignment of open large language models on high-stakes decision dilemmas, revealing a critical robustness gap in their behavioral safety without necessarily altering their underlying stable goals.
This paper proposes HACI-Net, a ConvNeXt-based hybrid network integrating spatial, coordinate, and channel attention mechanisms with channel interaction modules to effectively address challenges in food image recognition, achieving state-of-the-art accuracy on the VireoFood-172 and ChineseFoodNet datasets.
This paper presents a deep-learning framework based on a transformer architecture that analyzes 320 million matches from League of Legends to predict winning team compositions by jointly modeling individual player proficiency and team role congruency, significantly outperforming previous approaches.
This study demonstrates that selecting different large language models for anti-money laundering transaction screening leads to drastic variations in false positive rates and operational workloads, establishing model choice as a critical operational risk parameter that requires explicit governance and monitoring.
This paper proposes TriAdaptNER, a frequency-aware mixture-of-experts framework that leverages specialized high- and low-frequency experts guided by an adaptive selector and a differentiable error-driven feedback mechanism to effectively address class imbalance and improve recognition performance for rare entities in Named Entity Recognition tasks.
This paper proposes MOTMER-VANET, a multi-objective framework for Vehicular Ad Hoc Networks that integrates BiLSTM-based mobility prediction, reinforcement learning-driven trust management, and a hybrid optimization algorithm to dynamically select energy-efficient cluster heads and ensure secure, reliable routing against attacks.
This paper proposes G-Mamba, a gray-box multi-objective prediction network that integrates mechanistic knowledge with a bidirectional Mamba architecture and graph convolutional networks to achieve real-time, computationally efficient, and accurate forecasting of biomass and product concentrations in tyrosine fermentation processes by effectively capturing both long-sequence temporal dynamics and complex spatial couplings.
This paper introduces DACANet, a dual-path network that employs a learned adaptive gating mechanism to dynamically balance local convolutional and global transformer features, thereby achieving robust and generalizable medical image classification across diverse diagnostic domains without dataset-specific tuning.