3031 papers
This paper proposes SPQ-DETR, a transformer-based detector that enhances long-range tiny-UAV detection by integrating high-resolution feature preservation, prior-guided queries initialized from encoder tokens, and shape-aware geometric supervision to significantly improve accuracy and recall on challenging datasets.
This paper introduces \methodfull, a simulation-driven self-improving framework that iteratively generates, verifies, and repairs circuit netlists using Pareto-Verified Preference Optimization and Safe Proximal Policy Optimization to achieve high-precision, multi-objective point-spec circuit design across diverse analog families.
This paper proposes a robust EEG-based biometric identification system that converts 64-channel signals into 2D latent neural signatures via wavelet scalograms and a convolutional autoencoder, then classifies subjects using a custom ResNet to achieve 99.1% accuracy on the PhysioNet Motor Imagery dataset.
This paper proposes two Reservoir Computing-based architectural modifications that replace traditional, computationally complex MFCC feature extraction with efficient time-domain convolutions, achieving superior performance in speech recognition tasks through either a decentralized multi-reservoir ensemble or a single composite waveform approach.
This study utilizes reverse-correlation methods to demonstrate that multimodal LLMs like ChatGPT and Gemini generate stable, self-recognizing, and positively valenced "personified" facial images, providing preliminary evidence that these models may possess internal representations of their functional identity.
This study demonstrates that a Support Vector Machine (SVM) classifier, optimized using eight key features selected from the Cleveland dataset, achieves superior performance with 90.28% accuracy and a 93% AUC in predicting heart ailments compared to thirty other classifiers.
Over a two-year period, UNC Health enhanced its Responsible AI framework by implementing a three-tier risk-based system that significantly improved evaluation efficiency and reduced lead times while maintaining rigorous oversight, resulting in the successful review of 64 solutions and the establishment of robust post-deployment monitoring mechanisms with minimal adverse events.
This paper challenges the premise that active re-partitioning is universally necessary for vector-search indices under data drift, demonstrating through controlled benchmarks that static partitions suffice for moderate turnover while revealing that incremental re-centering is the cost-effective solution for significant distribution shifts, ultimately providing a regime map and decision rule for practitioners to determine when maintenance is truly required.
This paper presents a multi-agent reinforcement learning framework for safe industrial robot coordination, demonstrating through simulation that while a deterministic greedy baseline outperforms the proposed safe MARL policy in efficiency, the latter enables emergent communication patterns and highlights the critical need for adaptive communication bandwidth to mitigate collision risks as fleet sizes scale.
This paper demonstrates that a decision-aware machine learning framework for academic early warning outperforms traditional rule-based and probability-threshold approaches by reducing missed interventions and narrowing equity disparities through context-sensitive support decisions derived from calibrated risk estimates.