💻 Category

cs

2787 papers

💻 computer science

Leakage-Resistant Evaluation of Calibrated Multimodal Driver Drowsiness Detection Across Drivers and Routes

This study demonstrates that rigorous leakage-resistant protocols, particularly route-held-out validation, are essential for credible driver drowsiness detection, revealing that while pooled histogram-based gradient boosting achieves high performance under standard splits, it suffers a significant route-transfer penalty and that complex modality fusion does not outperform simpler pooled models.

J Robert Theivadas, Suresh Ponnan, Rinu Dhanaraj, Ruchi Patel2026-08-14
💻 computer science

Pixel-Translation-Equivariant Quantum Convolutional Neural Networks via Fourier Multiplexers

This paper introduces Pixel-Translation-Equivariant Quantum Convolutional Neural Networks (PCS-QCNNs) that resolve the mismatch between image encoding symmetries and standard qubit permutations by constructing Fourier-multiplexed layers that exactly commute with pixel cyclic shifts, demonstrating superior performance over non-equivariant quantum controls on translated MNIST benchmarks while highlighting critical train-deploy mismatches arising from finite-shot sampling costs.

Dmitry Chirkov, Igor Lobanov2026-08-14
💻 computer science

SPS-LIME: Role-Aware Segment Perturbation for Explaining Long Legal Documents

This paper introduces SPS-LIME, a model-agnostic explanation framework that enhances the interpretability of legal text classifiers by perturbing variable-length, rhetorically coherent segments rather than individual tokens or sentences, thereby achieving superior consistency and comprehensiveness while revealing potential mismatches between classifier sensitivity and explicit legal reasoning.

Meng-Luen Wu, Bo-Xun Huang2026-08-14
💻 computer science

Spatial Non-Stationarity and Wealth Accumulation: Multiscale Regression of Locational Endowments in a Heterogeneous Metropolitan County

This study introduces a hybrid GeoAI and Multi-Criteria Decision Analysis framework that utilizes unsupervised machine learning to identify legally and environmentally viable data center sites within the Guadalupe River Basin, effectively resolving water-energy conflicts by replacing subjective expert weighting with empirically derived criteria to guide infrastructure away from the sensitive Edwards Aquifer recharge zone.

Onyedikachi Joshua Okeke, Samuel Abakah, Oluwatosin Olofinsao, Sharma Suraj, Denis Baidoo, Emmanuel Ekubuafor, Olumide O (…)2026-08-14
💻 computer science

Token-Adaptive LoRA: Enhancing Segment Anything for Remote Sensing Imagery through Parameter-Efficient Fine- Tuning

This paper proposes Token-Adaptive LoRA, a parameter-efficient fine-tuning method that dynamically routes image tokens to rank-differentiated LoRA experts via a Noisy Top-1 router, significantly enhancing the Segment Anything Model's performance on remote sensing detection and segmentation tasks with minimal computational overhead.

Xin Chen, Jun Yan, Zhiyu Yan, Jianwen Deng, Jiaqi Wu, Yonghong Gong, Xiaohua Jiang2026-08-14
💻 computer science

Whose Art Counts? Model- and Prompt-Dependent Associations in Vision-Language Judgments of Museum Art

This paper introduces an archive-conditioned audit protocol to evaluate how vision-language models and prompt structures interact with museum metadata when assessing artistic value, demonstrating through a Metropolitan Museum of Art case study that "influence" prompts yield the most consistent category-based differences while highlighting the limitations of cross-model and cross-prompt generalization.

Manpreet Singh, Nandakishor Reddy Pulagam, Rhythm Bhatia, Rahul Joshi2026-08-14
💻 computer science

Improving Credit Card Fraud Detection Using Ensemble Machine Learning Techniques

This paper proposes a robust credit card fraud detection framework that combines Generative Adversarial Networks (GANs) to address severe class imbalance with ensemble machine learning techniques, demonstrating that the resulting XGBoost model achieves superior performance in recall, precision, and real-time feasibility on a dataset of 76,900 transactions.

Adel EL QASSOUARI, Ayoub CHAREF, Zahi JARIR2026-08-14
💻 computer science

A Query-Centric Diagnostic Attention BLSTM Network for Short-Window ECG-Based Atrial Fibrillation Detection

This study proposes a Query-Centric Diagnostic Attention BLSTM (QDA-BLSTM) network that leverages learnable diagnostic queries to capture complementary temporal information, demonstrating significantly improved accuracy and specificity over traditional CNN-BLSTM models for detecting atrial fibrillation in short-window ECG signals.

Yang Li, Jianguo Chen, Manhong Shi2026-08-14