AI Infrastructure
AI infrastructure and compute ecosystem
Articles
Event-Driven Reinforcement Learning Tackles Long-Horizon Fab Control⭐8
Event-driven reinforcement learning enables efficient long-horizon semiconductor fab control by acting on discrete events, reducing complexity and
Chip Industry Week In Review - Jun 19, 2026⭐7
Chip demand surges, AI hardware advances, packaging innovations, supply chain shifts, and 3nm production ramp define this week's semiconductor industry roundup.
Designing Chips That Can Explain Themselves⭐9
New chip designs embed built-in explainability, using trace buffers and dedicated engines to provide real-time, human-readable justifications for AI decisions.
Google Details Five Generations Of TPU Training Supercomputers⭐9
Google details five generations of TPU supercomputers, from v2 to v5p, highlighting architecture, performance, and AI training advances.
Chip Industry Technical Paper Roundup: June 16, 2026⭐7
Weekly roundup of semiconductor industry technical papers covering chip design, AI, architectures, and automotive advances, with links to full papers.
Physical Neural Networks: A Comprehensive Survey by University⭐9
University of Lübeck and TU Hamburg survey physical neural networks (PNNs) including in-memory computing and optical architectures for energy-efficient AI hardw...
Modeling Multi-GPU Traffic for Distributed AI Workloads |⭐9
UW-Madison and AMD model multi-GPU traffic to optimize distributed AI workloads, addressing communication bottlenecks in large-scale systems.
Wi-Fi Flies Higher as Edge AI Build-Out Takes Root⭐9
Wi-Fi 6E/7 and Edge AI converge in 2026, enabling low-latency, high-bandwidth connectivity for autonomous vehicles, smart factories, and AR systems.
Cross-Validated Timing Analysis for Automotive CAN Networks⭐9
A cross-validated timing analysis method improves CAN bus reliability in automotive networks, combining analytical models with real-world traffic logs for tight...
Refining Vision-Language Models for Lithography Defect Detection⭐9
Vision-language models enhance lithography defect detection by fusing image analysis with text for zero-shot classification, interactive querying, and