2731 papers
This paper introduces VeritasAI, an explainable multi-agent framework that leverages Retrieval-Augmented Generation and a graph-based knowledge store to detect fake news in real-time by simulating adversarial debate among specialized agents, thereby providing transparent, evidence-based verdicts and reasoning that outperform traditional black-box systems.
This study demonstrates that while a compact hybrid quantum-classical policy can marginally outperform a size-matched classical network in demand-charge-aware battery control, the advantage is narrow, task-specific, and does not constitute a general quantum advantage over faster classical baselines.
This paper introduces Recursive Inference Machines (RIMs), a neural reasoning framework that unifies neural backbones with classical recursive inference schemes to enhance performance on complex reasoning benchmarks and improve robustness and efficiency in tabular data settings.
This paper presents and validates a production-grade architecture for safe natural-language access to legacy enterprise systems that ensures reliability by replacing free-form SQL generation with a Schema-Metadata Navigation Graph and catalogue-constrained extraction, thereby eliminating hallucinations and security risks through architectural design rather than model scale.
This study presents a real-time intelligent monitoring system for microplastics that combines an improved data strategy with the YOLOv11 architecture to significantly enhance detection accuracy, reduce operator fatigue, and accelerate the screening process for complex microscopic samples.
Using Design Science Research, this study proposes and validates an AI-augmented agile framework that significantly accelerates MVP development and enhances quality for resource-constrained lean tech startups by reducing estimation time by 50% and achieving up to 97% migration accuracy through AI-assisted workflows.
This paper proposes a self-supervised, uncertainty-aware framework that improves object-level 3D localization stability in autonomous driving by leveraging temporal geometric consistency and ego-motion compensation to refine depth estimates without requiring costly 3D bounding-box annotations.
This paper proposes a comprehensive framework leveraging advanced machine learning and deep learning algorithms, including BERT and various RNN variants, alongside diverse embedding techniques and LDA topic modeling, to analyze patient textual feedback from 21 hospitals and derive actionable insights for enhancing healthcare services and patient satisfaction.
This survey reviews foundation model-based industrial visual anomaly detection under a cold-start deployment framework, systematically categorizing methods by data availability and pipeline roles while addressing challenges in adaptation, evaluation, and scalable implementation.
WiFlux AI is an intelligent framework that optimizes smart campus Wi-Fi by using real-time analysis of 14 parameters per access point to autonomously predict and resolve congestion, achieving significant improvements in network speed, packet loss, and congestion reduction while scaling to support up to 1,000 access points.