Research & Innovation

Global: New Framework Tackles Multi-Model Training Challenges in Vehicle-Edge-Cloud Federated LearningResearchers announced a novel framework on arXiv on January 31, 2025, aimed at improving hierarchical federated learning across vehicle‑edge‑cloud (VEC)…

Global: LLMs Guide Compiler Optimizations in Closed-Loop FrameworkResearchers have introduced a new framework that uses off‑the‑shelf large language models (LLMs) to guide code optimization for complex loop nests. The system…

Global: Markovian Reasoning Framework Reduces Redundant Computation in LLM Inference Researchers have introduced a Markovian reasoning process designed to reduce redundant computations in large language model (LLM) inference, according to…

Global: Researchers Unveil Integrated Adaptive Design and Bayesian Inference FrameworkResearchers have introduced a novel framework that jointly amortizes adaptive experimental design and Bayesian inference, aiming to streamline the process of…

Global: New Neural ODE‑Based Framework Improves Multimodal Image RegistrationResearchers announced a novel multimodal diffeomorphic registration technique in a December 2025 arXiv preprint. The method leverages Neural Ordinary Differential Equations to…

Global: Study Generates Diverse Near-Optimal Paths for European Green E‑Molecule ImportsResearchers Mahdi Kchaou, Francesco Contino, and Diederik Coppitters posted a new preprint on arXiv on 29 December 2025 that explores…

Global: FedORA Enables Efficient Unlearning in Vertical Federated LearningA team of machine learning researchers announced a new algorithm designed to remove specific data influences from models trained under vertical federated…

Global: Warning-Framed Training Data Does Not Significantly Reduce Undesired Model Outputs, Study FindsOn December 25, 2025, a study authored by Tsogt-Ochir Enkhbayar reported that language models exposed to warning-framed training…

Global: Researchers Map Growing Landscape of Threats to Foundation AI ModelsResearchers affiliated with multiple institutions released a new pre‑print on arXiv on December 2025 that outlines a comprehensive assessment of…

Global: SemCovert Framework Introduces Semantic-Level Hiding for Secure Video TransmissionA preprint posted on arXiv (ID 2512.22233) presents SemCovert, a deep learning framework designed to embed secret video content within semantic…