Infrastructure & Hardware

B
Showing 121–149 of 149
  • Data Center Dynamics · EN Infrastructure & Hardware
    Sponsored: Making data centers ready for AI workloads with rack-level cooling
    Making data centers AI-ready with rack-level cooling (sponsored)
    As AI workloads mature and scale, organizations must decide where high-density compute will run. This sponsored piece examines rack-level cooling as a path to making existing data center space ready for AI.
    Read original (Data Center Dynamics) ↗
  • The Register (Data Centre) · EN Infrastructure & Hardware
    Nvidia's Groq acquihire is on the DOJ's radar, but it's already too late
    Nvidia's $20B Groq acquihire draws DOJ scrutiny, likely too late
    NVIDIA
    The DOJ is reportedly examining Nvidia's roughly $20 billion acquihire of Groq. The Register argues that unwinding the deal would change little, since a growing list of inference-chip alternatives is already positioned to fill Groq's place.
    Read original (The Register (Data Centre)) ↗
  • Hacker News (Front Page) · EN Infrastructure & Hardware
    The EPA Is Planning to Scrap Public Review Rules for Data Center Pollution
    EPA moves to end public comment on data center air permits
    Neural Network
    The US EPA proposed scrapping a federal rule requiring states to notify the public and take comments before issuing air-pollution permits, covering data centers and the power plants feeding them. About 200 advocacy groups and a dozen-plus states oppose it.
    Read original (Hacker News (Front Page)) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN Inference & Efficiency
    Rethinking Heterogeneous System Disaggregation for Subquadratic Attention
    Inference Retrieval-Augmented Generation (RAG)
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN Infrastructure & Hardware
    Embodied-BenchForge: A Closed-Loop Agentic Workflow for Embodied Benchmark Construction
    AI Agents Machine Learning
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • arXiv cs.CL (Computation and Language) · EN New Model Releases
    Expert-Space Exploration in MoE Reinforcement Learning
    Mixture of Experts (MoE) Reinforcement Learning
    Read original (arXiv cs.CL (Computation and Language)) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN Infrastructure & Hardware
    Diffusion Models and Concept Formation
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • arXiv cs.CL (Computation and Language) · EN Infrastructure & Hardware
    Kraken: LLM-based Speech-to-Speech Translation via Low-bitrate VQ and Dual-path Source Conditioning
    Machine Learning Quantization Speech Processing
    Read original (arXiv cs.CL (Computation and Language)) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN Infrastructure & Hardware
    DynSHAP: Towards Explainable Dynamic Survival Analysis
    Deep Learning Neural Network Reinforcement Learning
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • Data Center Dynamics · EN Infrastructure & Hardware
    Plans for 120MW data center withdrawn in Lombardy, Italy
    120MW data center plan withdrawn in Lombardy, Italy
    Gemini Neural Network
    Plans for a 120MW data center in Lombardy, Italy have been withdrawn. The local mayor said the municipality never received the data it needed for the grid connection, making power-infrastructure requirements the deciding factor in the project's fate.
    Read original (Data Center Dynamics) ↗
  • arXiv cs.CL (Computation and Language) · EN New Model Releases
    Investigating Temporal Motion Features for Pose-to-Text Indian Sign Language Translation
    Embeddings
    Read original (arXiv cs.CL (Computation and Language)) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN New Model Releases
    Generative Retrieval for Unsupervised Text-Based Person Search
    Neural Network Retrieval-Augmented Generation (RAG) Software Engineering
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN Multimodal
    Tracing and Coordinating Cross-Layer Influence for Multimodal Model Merging
    Reinforcement Learning
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • Simon Willison's Weblog · EN New Model Releases
    Don't sleep on wrapture
    Dumpleton's wrapture unifies Python mocking and tracing
    Graham Dumpleton released wrapture, a Python monkey patching library serving both unit testing and New Relic-style observability. Near-daily tutorials since its August 31 debut cover call recording as trees, phased behaviour across successive calls, and patching attributes, dictionaries and generators.
    Read original (Simon Willison's Weblog) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN Infrastructure & Hardware
    4D Parallelism Unlocks Exascale Bayesian Neural Networks for High-Fidelity Atmospheric Modeling
    Neural Network NVIDIA Retrieval-Augmented Generation (RAG) Reinforcement Learning Transformer
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN New Model Releases
    K-Bench: A Benchmark for LLM Unlearning in Agentic Deployments
    Neural Network Software Engineering
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN New Model Releases
    Interpreting the predictions of neural network classification based on a Taylor Coefficient Analysis (TCA)
    Neural Network
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • Data Center Dynamics · EN Infrastructure & Hardware
    Orbital partners with Reflex Aerospace for space data center constellation
    Orbital partners with Reflex Aerospace on space data center constellation
    Orbital has partnered with Reflex Aerospace to build a constellation of data centers in space. The deal gives the experienced German developer its first US customer, and Reflex has promised to supply its technology to Orbital exclusively.
    Read original (Data Center Dynamics) ↗
  • arXiv cs.AI (Artificial Intelligence) · EN Infrastructure & Hardware
    Unified Agentic Video Editing Across Levels of Complexity and Creativity
    Neural Network
    Read original (arXiv cs.AI (Artificial Intelligence)) ↗
  • Data Center Dynamics · EN Infrastructure & Hardware
    Microsoft targets 38GW of data center capacity in 2032 - report
    Microsoft targets 38GW of data center capacity by 2032, report says
    Microsoft
    Microsoft is reportedly targeting 38GW of data center capacity by 2032. The company currently operates around 12GW, so the goal would mean more than tripling its footprint over roughly six years.
    Read original (Data Center Dynamics) ↗
  • Data Center Dynamics · EN Infrastructure & Hardware
    Oracle delivered 300,000 GPUs in Q1 FY2027
    Oracle delivered 300,000 GPUs in Q1 FY2027
    Oracle said it delivered 300,000 GPUs in the first quarter of its 2027 fiscal year. Executives also teased a "new agentic AI accelerator," and the figure points to a sharp expansion in the compute capacity the cloud provider is supplying.
    Read original (Data Center Dynamics) ↗
  • arXiv cs.CL (Computation and Language) · EN Infrastructure & Hardware
    Residual Vector-based Reconstruction as Long-Context Recall Regardless of Context Window Size
    Deep Learning Fine-tuning Software Engineering
    Read original (arXiv cs.CL (Computation and Language)) ↗
  • Data Center Dynamics · EN Infrastructure & Hardware
    Amid data center public backlash, Yvette Eden Ruiz joins OpenAI’s Community Engagement team from JPMorgan
    OpenAI hires JPMorgan's Yvette Eden Ruiz for community engagement
    OpenAI
    OpenAI has hired Yvette Eden Ruiz from JPMorgan for its Community Engagement team. The move comes as public backlash against data centers grows into an issue heading into the midterms, with OpenAI looking to reshape how it communicates.
    Read original (Data Center Dynamics) ↗
  • arXiv cs.CL (Computation and Language) · EN Multimodal
    Calibrated Ambiguity in Multimodal Language Models: Humans reach for cultural references, while models describe the picture
    Reinforcement Learning
    Read original (arXiv cs.CL (Computation and Language)) ↗
  • arXiv cs.CL (Computation and Language) · EN Infrastructure & Hardware
    The House with a Million Windows: Interactive Fiction for Narrative Restorying
    Read original (arXiv cs.CL (Computation and Language)) ↗
  • The Register (Data Centre) · EN New Model Releases
    DeepSeek's new model sets a template for powerful LLMs that run lean
    DeepSeek V4.1 Flash sets a template for powerful LLMs that run lean
    DeepSeek Machine Learning
    DeepSeek's new V4.1 Flash model shows that a bigger model does not necessarily require more GPUs to serve. The Register frames it as a template for powerful LLMs that can run lean, decoupling model scale from serving footprint.
    Read original (The Register (Data Centre)) ↗
  • arXiv cs.CL (Computation and Language) · EN Infrastructure & Hardware
    Confidence-Gated Transductive Test Generation for Code Reranking
    Read original (arXiv cs.CL (Computation and Language)) ↗
  • arXiv cs.CL (Computation and Language) · EN Infrastructure & Hardware
    AMDKernelVault: Large-Scale Datasets and Agentic Training for AMD GPU Kernel Optimization
    AI Agents Fine-tuning Machine Learning NVIDIA Reinforcement Learning
    Read original (arXiv cs.CL (Computation and Language)) ↗
  • arXiv cs.CL (Computation and Language) · EN Infrastructure & Hardware
    Representation-based Masked Diffusion Model
    Retrieval-Augmented Generation (RAG) Reinforcement Learning
    Read original (arXiv cs.CL (Computation and Language)) ↗