4157 papers
AdaSeq-MTA is an adaptive neural framework that significantly reduces computational latency for multi-touch attribution by dynamically selecting touchpoints and adjusting attribution effort while maintaining predictive accuracy and minimizing credit allocation drift through a teacher-guided, workload-constrained optimization approach.
This systematic mapping study analyzes 1,252 research papers on explainability methods for large language models, revealing that most techniques require inaccessible internal model data and that evidence regarding human utility and real-world deployment remains critically scarce.
This scoping review of 71 studies published between 2023 and 2025 reveals that while large language models are increasingly applied to clinical trial recruitment and data analysis, their transition to routine practice is hindered by fragmented evidence, a lack of comparative benchmarks, and insufficient prospective validation.
This paper proposes a novel non-data-driven hierarchical framework called NDVM for general video matting that achieves superior temporal coherence and accuracy by sequentially employing KNN-based pixel classification, optical-flow-based sampling with re-identification, and affinity-based smoothing without relying on fixed datasets.
This bibliometric analysis of 646 meta-analyses on AI in healthcare (2010–2025) reveals a dramatic post-2019 surge in publications led by China and the UK, highlighting rapid growth that has outpaced methodological consolidation and international institutional collaboration.
This study demonstrates that implementing event-triggered activity notifications significantly enhances collective information recovery in large language model teams following communication disruptions, increasing success rates from 2.5% to 76.25% compared to teams relying solely on agent-initiated checks.
This paper proposes an improved YOLO model for mandibular fracture detection in CT images that integrates a novel Cross-directional Regional Pooling Attention (CRPA) module to enhance spatial dependency modeling and positional information preservation, achieving superior performance metrics compared to baseline and state-of-the-art methods.
This survey presents a comprehensive framework for empowering the Internet of Vehicles with embodied intelligence by systematically analyzing the closed-loop integration of perception, decision-making, and control, while evaluating existing technologies, datasets, metrics, and future challenges to enhance vehicle adaptability and safety.
This paper proposes Redundancy-Triggered Relocation (RTR), a decentralized coordination mechanism where local searchers autonomously restart upon detecting fitter teammates within a limited perception radius, effectively reducing redundant convergence and significantly improving performance on dynamic multimodal optimization benchmarks without requiring a central controller or problem-specific parameter tuning.
HSR-APR is a hybrid automated program repair framework that integrates static analysis, security validation, and an Adaptive Reflection Memory mechanism to significantly improve the success rate, security, and efficiency of LLM-based patch generation compared to baseline approaches.