Research & Innovation

Global: New Benchmark Highlights LLMs’ Sensitivity to Instruction OrderOverview of the RIFT TestbedA new benchmark called the Reordered Instruction Following Testbed (RIFT) has been introduced to evaluate how large language…

Global: New Multi-Language Benchmark Targets Neural Theorem Proving for Software Verification ConditionsLeadResearchers from multiple institutions have released the first multi-language benchmark designed to evaluate neural theorem proving on verification conditions…

Global: New Uncertainty-Aware Model Boosts Emotional Talking Face SynthesisResearchers have unveiled UA-3DTalk, an uncertainty-aware 3D emotional talking-face synthesis framework that aims to improve audio-vision emotion alignment and adaptive multi-view fusion,…

Global: LLM Agents Enable Autonomous Knowledge Discovery in Atomic Layer Processing SimulationsA study released on arXiv on September 30, 2025, and revised on January 27, 2026, demonstrates that large language…

Global: Room Environment v3 Introduces Temporal Knowledge Graph Memory for Agent ResearchResearchers behind a recent arXiv preprint have unveiled a configurable simulation called the Room Environment v3, designed to test…

Global: Study Evaluates Robustness of Time-Series AI Models Using Causal Rating FrameworkA new study released in February 2025 introduces a causally grounded rating framework designed to evaluate the robustness of…

Global: New Task Allows AI Models to Incorporate New Table Columns During InferenceA team of machine learning researchers has introduced a novel task designed to let trained models adapt to…

Global: Study Finds Voice-Interactive Agent Aligns With Specialists in Detecting Alzheimer’s SymptomsResearchers have evaluated a voice-interactive conversational agent powered by large language models to gather patient narratives relevant to early…

Global: New Mamba-Based Framework Boosts Multi-Modal Video Temporal GroundingResearchers Zhiyi Zhu, Xiaoyu Wu, Zihao Liu, and Linlin Yang released a preprint on arXiv on June 10, 2025, with a revised…

Global: AI‑Based TELGEN Algorithm Cuts Traffic Engineering Time While Maintaining Near‑Optimal PerformanceResearchers have introduced TELGEN, an artificial intelligence–based traffic engineering (TE) algorithm designed to handle the scalability demands of extensive…