927 papers
This paper introduces UD-Quran, a Universal Dependencies v2 conversion of the Extended Quranic Treebank that standardizes morphological and syntactic annotations into two interoperable variants (surface and augmented) to enable consistent Quranic parsing with modern NLP tools while preserving traceability to the original corpus.
This paper introduces TRNEWS-2025, a large-scale, manually annotated Turkish news dataset covering nine diverse categories up to 2025, which has been validated for effective use in text classification and NLP research through experiments with both transformer-based and classic deep learning models.
This paper proposes MM-QLOF, a decentralized multi-metric Q-learning objective function that integrates packet success rate, expected transmissions, and congestion levels to enable autonomous, adaptive parent node selection in IoT networks, significantly improving packet delivery rates and reducing latency compared to existing RPL standards.
This paper presents a conceptual narrative review of literature from 2020 to 2025 on multi-case global cyberattacks, highlighting the limitations of traditional security models and proposing a new four-tier framework to advance the understanding and implementation of cyber resilience.
This paper proposes a multi-sensor privacy data tampering prevention method that combines distributed authentication, an improved AES-encrypted T-box, and differential variance-based flood attack detection to ensure the security, integrity, and confidentiality of data during transmission and storage.
This paper proposes FPSD-MCP, a novel federated learning framework that integrates secure multi-party computation and homomorphic encryption with an adaptive aggregation strategy to achieve a scalable, privacy-preserving, and communication-efficient spam detection solution for heterogeneous multi-cloud environments.
This paper proposes the Intent-Guided Diffusion Model (IGDM), a novel framework that enhances user cold-start recommendation by extracting transferable latent intent prototypes from historical data, adaptively fusing them with sparse user behaviors, and guiding a diffusion-based denoising process to refine preference estimation, thereby outperforming existing baselines on multiple real-world datasets.
This paper introduces TRIDIS, an open, harmonized corpus of medieval and early modern handwritten documents designed to support Handwritten Text Recognition and Named Entity Recognition, while proposing an innovative outlier-based evaluation strategy to address the limitations of conventional random splits in heritage-document benchmarking.
This paper introduces the Aye-Aye Optimizer (AAO), a novel bio-inspired metaheuristic algorithm that mimics the Aye-Aye lemur's percussive foraging strategy through three adaptive phases to achieve superior balance between exploration and exploitation, faster convergence, and robust performance on benchmark and real-world optimization problems.
This paper proposes a vision language model-based framework with constraint example verification feedback and confidence evaluation to achieve high-accuracy structured extraction and standardized reconstruction of heterogeneous machining process documents under low-resource constraints, effectively reducing hallucinations and enabling knowledge digitization in intelligent manufacturing.