Infrastructure & Hardware B

Showing 61–90 of 217
  • arXiv cs.AI (Artificial Intelligence) · EN Inference & Efficiency
    Towards Autonomous Aircraft Surveillance from Nanosatellites through On-Board Inference and Generative Data Augmentation
    Inference Neural Network Retrieval-Augmented Generation (RAG)
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN New Model Releases
    SVR: Self-Verifying Refinement via Joint Verdict-Confidence Reinforcement Learning for Adaptive Test-Time Compute
    Inference Retrieval-Augmented Generation (RAG) Reinforcement Learning Software Engineering
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN Inference & Efficiency
    WIDE: Boosting Adaptive LLM Inference via Token-level Dynamic Width Pruning
    Deep Learning Inference Neural Network Reinforcement Learning
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • NVIDIA Developer Blog · EN Infrastructure & Hardware extract
    NVIDIA Exemplar Cloud: Lessons for Unlocking Full Performance on AI Infrastructure
    NVIDIA shares Exemplar Cloud lessons for unlocking AI infra performance
    NVIDIA
    NVIDIA shared lessons from its Exemplar Cloud, noting that two clusters built from identical H100, GB200 NVL72, or GB300 NVL72 systems can deliver materially different performance. The guidance focuses on tuning and operations to unlock full AI infrastructure performance.
    Read original (NVIDIA Developer Blog) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN Infrastructure & Hardware
    When Specifications Conflict: A Symmetry-Based Framework for Measuring LLM Preferences
    Neural Network
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN Infrastructure & Hardware
    How Benchmarks Mis-Score Computer-Use Agents
    AI Agents Neural Network
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • Data Center Dynamics · EN Infrastructure & Hardware extract
    New Jersey town sued for $300m by data center developer
    Data center developer sues New Jersey town for $300m
    Meta Reinforcement Learning
    A data center developer has sued a New Jersey town for $300m, alleging the town unfairly rejected its proposal, according to DatacenterDynamics. The case is another example of disputes between data center builders and local governments escalating into litigation.
    Read original (Data Center Dynamics) ↗
  • arXiv cs.CL (Computation and Language) · EN New Model Releases
    LLMs struggle to simulate human belief updates in controlled environments
    GPT
    Read original (arXiv cs.CL (Computation and Language)) ↗
  • Hugging Face Blog · EN Infrastructure & Hardware extract
    GPU Management: Why Idle GPUs Are the New Grounded Aircraft
    Hugging Face: idle GPUs are the new grounded aircraft
    Deep Learning
    Hugging Face likened idle GPUs to grounded aircraft, arguing that unused accelerators represent a major cost drain. The piece stresses better GPU management and scheduling to raise utilization and avoid expensive idle capacity.
    Read original (Hugging Face Blog) ↗
  • Data Center Dynamics · EN Infrastructure & Hardware extract
    Sponsored: Four common misconceptions about nickel-zinc battery technology
    Four common misconceptions about nickel-zinc batteries (sponsored)
    A sponsored article addresses four common misconceptions about nickel-zinc battery technology. It argues that the choices organizations make today will shape not only performance and reliability but also how the next generation of infrastructure is built.
    Read original (Data Center Dynamics) ↗
  • arXiv cs.CL (Computation and Language) · EN Funding & M&A
    Fairness Pruning: Locating Demographic Bias in GLU-MLP Layers via Differential Activations
    Inference Llama Machine Learning Meta
    Read original (arXiv cs.CL (Computation and Language)) ↗
  • Data Center Dynamics · EN Infrastructure & Hardware extract
    Crusoe partners with Aalo on nuclear-powered AI data center deployment at Idaho National Lab
    Crusoe teams with Aalo on SMR-powered AI data center at Idaho National Lab
    Crusoe is partnering with Aalo to deploy a nuclear-powered AI data center at Idaho National Lab, DatacenterDynamics reported. Crusoe will install a Spark unit at the site, which is to be powered by a small modular reactor (SMR), reflecting growing interest in nuclear for AI compute.
    Read original (Data Center Dynamics) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN New Model Releases
    ObjectStream: Latent Objects as Memory Anchors for Streaming Video Understanding
    Reinforcement Learning
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • arXiv cs.LG (Machine Learning) · EN Inference & Efficiency
    Fully Inductive Cardinality Estimation
    Embeddings Neural Network Reinforcement Learning
    Read original (arXiv cs.LG (Machine Learning)) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN Infrastructure & Hardware
    MonoVoc: Decoupling Geometry and Semantics for Lightweight Monocular Open-Vocabulary 3D Gaussians
    Embeddings Retrieval-Augmented Generation (RAG) Software Engineering
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • Data Center Dynamics · EN Infrastructure & Hardware extract
    Vertiv's earnings and revenues grow, shares fall
    Vertiv grows earnings and revenue, but shares fall
    Vertiv reported higher earnings and revenue, yet its shares fell, DatacenterDynamics reported. The decline comes as AI data center investment grows increasingly volatile, with investors wary of uncertainty despite the power and cooling supplier's solid results.
    Read original (Data Center Dynamics) ↗
  • arXiv cs.CL (Computation and Language) · EN New Model Releases
    (Towards) Scalable Reliable Automated Evaluation with Large Language Models
    Retrieval-Augmented Generation (RAG)
    Read original (arXiv cs.CL (Computation and Language)) ↗
  • Data Center Dynamics · EN Infrastructure & Hardware extract
    Crusoe files for two more data centers at Goodnight Campus in Armstrong County, Texas
    Crusoe files for two more data centers at Texas Goodnight Campus
    Crusoe filed to build two more data centers at its Goodnight Campus in Armstrong County, Texas, DatacenterDynamics reported. Buildings 3 and 4 are each expected to cost $500m, continuing the company's large-scale expansion of AI-focused capacity.
    Read original (Data Center Dynamics) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN Infrastructure & Hardware
    Agentic Method for Deterministic Validation of Legacy Code Migration
    Neural Network Retrieval-Augmented Generation (RAG) Reinforcement Learning
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • arXiv cs.LG (Machine Learning) · EN Safety & Evaluation
    Uncertainty quantification for trustworthy deep learning: Methods and measures
    Deep Learning Neural Network Reinforcement Learning
    Read original (arXiv cs.LG (Machine Learning)) ↗
  • Data Center Dynamics · EN Infrastructure & Hardware extract
    Al Moammar Information Systems to explore building three new data centers in Saudi Arabia
    Al Moammar to explore building three new data centers in Saudi Arabia
    Al Moammar Information Systems is exploring the construction of three new data centers in Saudi Arabia, DatacenterDynamics reported. A deal with Saudi Data Center Fund 1 would add another 48MW to the country's data center capacity, supporting its digital-infrastructure ambitions.
    Read original (Data Center Dynamics) ↗
  • Data Center Dynamics · EN Infrastructure & Hardware extract
    Energy firm Uniper to build new data center hub in Hesse, Germany
    Energy firm Uniper to build new data center hub in Hesse, Germany
    Retrieval-Augmented Generation (RAG)
    Energy company Uniper will build a new data center hub in Hesse, Germany, DatacenterDynamics reported. The complex at Großkrotzenburg will also host a new power plant and a battery energy storage system, pairing compute capacity with dedicated on-site power.
    Read original (Data Center Dynamics) ↗
  • IEEE Spectrum (AI section) · EN Inference & Efficiency extract
    Are AI Models Working Harder Than They Need to?
    Are AI models working harder than they need to?
    Deep Learning Google Inference Neural Network Software Engineering
    IEEE Spectrum examines how much of modern AI relies on massive amounts of multiplication. Questioning whether the neural networks behind everything from generated answers to photo organization really need all that computation, the piece explores the potential to make AI inference far more efficient.
    Read original (IEEE Spectrum (AI section)) ↗
  • arXiv cs.CL (Computation and Language) · EN Infrastructure & Hardware
    Fidelity Is Not Safety: Gently-Compressed LLMs Pass Every Data-Free Quality Guard Yet Invent Procedure Steps in Agentic Execution
    Machine Learning Neural Network
    Read original (arXiv cs.CL (Computation and Language)) ↗
  • Data Center Dynamics · EN Infrastructure & Hardware extract
    PowerPlay AI, Sharon AI plan 400MW behind-the-meter gas-powered data center in Abilene, Texas
    PowerPlay AI, Sharon AI plan 400MW behind-the-meter gas data center in Texas
    PowerPlay AI and Sharon AI plan a 400MW behind-the-meter, gas-powered data center in Abilene, Texas, DatacenterDynamics reported. The two companies have formed a joint venture to develop the facility, tapping on-site gas generation to meet heavy AI power demand.
    Read original (Data Center Dynamics) ↗
  • Data Center Dynamics · EN Infrastructure & Hardware extract
    Innio secures 1.1GW gas engine order from data center developer in US
    Innio secures 1.1GW gas engine order from US data center developer
    Neural Network
    Innio secured a 1.1GW gas engine order from a US data center developer, DatacenterDynamics reported. The engines will power data centers behind-the-meter, bypassing the grid—another sign that AI-driven demand is pushing operators to lock in dedicated power sources.
    Read original (Data Center Dynamics) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN New Model Releases
    Where and When to Commit: Candidate-Aware Decoding for Diffusion Language Models
    Computer Vision Deep Learning Reinforcement Learning Software Engineering
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN Safety & Evaluation
    Asymmetric Communication: Large Language Models and Language Games
    Neural Network
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • arXiv cs.LG (Machine Learning) · EN Training & Fine-tuning
    LM-GRASP: Instance-Specific Language Models for Combinatorial Construction via Online Imitation Learning
    Machine Learning Meta Retrieval-Augmented Generation (RAG) Reinforcement Learning Transformer
    Read original (arXiv cs.LG (Machine Learning)) ↗
  • Data Center Dynamics · EN Infrastructure & Hardware extract
    Beetaloo Energy partners with Halliburton to develop natural gas plant for proposed data center in Australia - report
    Beetaloo Energy, Halliburton to build gas plant for Australian data center
    Neural Network
    Beetaloo Energy is partnering with Halliburton to develop a natural gas plant for a proposed data center in Australia, according to a report cited by DatacenterDynamics. Up to 2GW of natural gas generation is planned for the site to power the facility.
    Read original (Data Center Dynamics) ↗