Research & Innovation

Global: Study Finds Large Language Models Enable Novice Users to Perform Complex Web ScrapingResearchers publishing a new paper on arXiv have demonstrated that large language models (LLMs) now allow users…

Global: Backdoor Threats Identified in Knowledge DistillationIn April 2025, researchers published a paper on arXiv that demonstrates a new class of backdoor attacks targeting the knowledge distillation process used to…

Global: Tokenizers Revealed as New Vector for Membership Inference Attacks on Large Language ModelsResearchers publishing on arXiv in October 2025 introduced tokenizers as a novel attack vector for membership inference…

Global: Study Evaluates Legal Admissibility of AI-Generated Forensic EvidenceIn a recent arXiv preprint, scholars assess whether artificial‑intelligence‑generated forensic evidence satisfies the reliability standards required for admission in criminal trials across…

Global: New Relative Disclosure Risk Indicator Helps Choose Differential Privacy ParametersA team of computer scientists has introduced a metric that quantifies how the privacy parameter ε influences the disclosure risk of…

Global: IsabeLLM Automates Formal Verification of Blockchain Consensus ProtocolsBackground on Consensus VerificationOn 12 January 2026, researchers Elliot Jones and William Knottenbelt posted a preprint titled Towards Automating Blockchain Consensus Verification…

Global: Researchers Propose Tighter Upper Bound for Shortest Vectors in Ideal LatticesIn a newly posted arXiv preprint (ID 2601.07511), a team of cryptography researchers introduced a novel analytical method for…

Global: Distillation May Increase Membership Inference Risks in Large Language ModelsA recent study posted on arXiv in May 2025 examines how knowledge distillation influences membership inference attacks (MIAs) on large…

Global: New Multimodal AI Framework Aims to Predict Cardiovascular Risk with Interpretability and PrivacyResearchers announced in January 2026 a comprehensive artificial‑intelligence system designed to forecast individual cardiovascular disease (CVD) risk…

Global: New Open-Source Library ‘skwdro’ Enables Wasserstein Distributionally Robust Machine LearningOn January 9, 2026, a team of computer scientists led by Florian Vincent released skwdro, a Python library intended to…