Large Language Model– tag –
-
Φ-Bench: New Benchmark with 85 Real-World Tasks Evaluates LLMs’ Ability to Engineer and Optimize Their Own Infrastructure
arXiv USA Overview Researchers have introduced Φ-Bench (Frontier AI Infrastructure Benchmark), a novel evaluation suite comprising 85 real-world tasks designed to assess Large Language Models' (LLMs) capacity to develop and optimize thei... -
NovGauge: A Human-Anchored Fine-Grained Benchmark Diagnoses LLMs’ Capability in Scientific Paper Novelty Assessment, Revealing Hallucinations and Lack of Justification
arXiv USA Overview Researchers have proposed NovGauge, a fine-grained, human-anchored benchmark designed to diagnose Large Language Models' (LLMs) capabilities in assessing scientific paper novelty. Comprising 619 paper pairs and 50 mult... -
TrendForce Report: Global Data Center Power Gap to Reach 268GW by 2030 Due to Surging AI Compute Demand
BigGo Finance Taiwan Overview Citing TrendForce, global power consumption is accelerating due to fierce AI computing competition, with data center power demand expected to hit 161GW in 2026, where AI servers will account for over one-thi... -
Large Language Models Autonomously Generate Chemical Synthesis Protocols, Bridging ‘Lab-to-Fab’ Gap in Materials Manufacturing
MDPI Switzerland Overview This review focuses on AI-driven strategies to bridge the 'lab-to-fab' gap in materials synthesis and manufacturing, moving beyond property prediction to address manufacturability. It discusses the use of Large ... -
Materials Discovery Bottleneck Lies in Model ‘Composition’ Capability: LLM Agents Key to Integrating Specialist Models
RSC Digital Discovery UK Overview The primary bottleneck in materials discovery is not the predictive accuracy of individual models but the lack of a layer to compose specialist models into context-dependent, multi-step scientific invest... -
Hugging Face Reveals LLM Agent-Driven Hypothesis Generation for Materials Discovery, Accelerating Solid-State Exploration with MatExpert
Hugging Face USA Overview A recent collection of research papers published by Hugging Face includes advancements in hypothesis generation for materials discovery and design using goal-driven and constraint-guided LLM agents. Notably, the... -
Multi-Agent LLM ‘MAESTRO’ Designs Single-Atom Catalysts via Reasoning, Breaking Conventional Limits
ACS Publications USA Overview MAESTRO (Multi-Agent-based Electrocatalyst Search Through Reasoning and Optimization) is a reasoning-driven framework where multiple LLMs collaboratively design high-performance single-atom catalysts. LLM ag... -
Large Language Model Agents Achieve 50-88% Success Rate in Biopharma Automated Synthesis, Accelerating DMTA Cycles
PharmExec USA Overview Agentic AI is revolutionizing the biopharmaceutical sector, with automated small molecule synthesis platforms like onepot.ai gaining prominence. These integrated systems, combining retrosynthesis engines, web searc... -
LLM Agents Unleash Autonomous Materials Discovery and Design
Unknown International Overview Large language model (LLM)-powered agents are now autonomously executing complex scientific workflows in computational materials science, significantly enhancing research efficiency and reliability. These a... -
Qualcomm and Amazon Forge Strategic Partnership for Custom AI Chips, Amazon Secures Warrants Tied to $60 Billion Purchase Agreement
BigGo Finance USA Overview Qualcomm and Amazon have announced a multi-generational strategic partnership to jointly develop custom AI inference chips and optical interconnect solutions for AI data centers. The agreement includes Qualcomm...
