3031 papers
This study presents a Design Science prototype of an Intelligent Learning Management System that leverages the ESCO ontology and NLP to detect workforce skill gaps in SMEs and generate targeted reskilling recommendations, thereby addressing the human capital–technology complementarity constraints critical to Industry 4.0 and 5.0 transitions.
This study demonstrates that an Ant Colony Optimisation approach incorporating flexible capacity constraints significantly outperforms a geographic clustering baseline in a real-world Colombian inbound logistics network, achieving substantial reductions in fleet size, travel distance, and operating costs while maximizing vehicle utilization.
This study demonstrates that naively concatenating demographic metadata with text inputs in a DistilBERT-based Automated Essay Scoring model significantly degrades predictive accuracy, increases training loss, and exacerbates scoring bias compared to a text-only baseline.
This paper introduces StainSolver, a method that accelerates the StainFuser diffusion model for histopathology stain normalization by replacing its PNDM scheduler with the efficient DPM-Solver++ and evaluating performance with a novel Stain Normalization Quality Score (SNQS), achieving significant speedups and improved quality without retraining.
FlashSpec is an open-source, adaptive speculative decoding engine that achieves exact target-model distribution preservation through a novel O(1) vocabulary-size Triton GPU kernel for on-device verification and an online bandit-based mechanism for dynamic draft-model selection, significantly accelerating LLM inference while eliminating CPU bottlenecks and manual tuning.
This paper proposes a hybrid unsupervised anomaly detection framework combining Autoencoders and Isolation Forests to effectively identify fake e-commerce reviews using TF-IDF and BERT features, offering a scalable and cost-effective alternative to supervised methods by eliminating the need for labeled training data.
This paper proposes a novel deep learning framework that integrates long-term dual-task motor and cognitive data with block-wise resampling, SMOTE augmentation, and a Transformer model to accurately predict two-year cognitive decline with 80% precision, outperforming traditional MRI and biomarker-based methods.
This paper proposes Trusted and Explainable Collective Learning (TECL), a novel method that actively leverages conflicts in multi-view data to assess individual consistency, prioritize views, and generate reliable, interpretable collective decisions, rather than simply eliminating such conflicts.
This paper proposes a computationally efficient, word-embedding-based data augmentation strategy using neighbor word replacement that significantly improves Named Entity Recognition and offers modest gains for Entity Linking across five languages, thereby addressing the challenge of limited annotated clinical text in multilingual health informatics.
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.