China’s Top Chip Lab Unveils Breakthroughs in Quantum Sensing, Solar Cells, and AI‑Inspired Brain Research

The Institute of Microelectronics (IME) at the Chinese Academy of Sciences has announced a wave of cutting‑edge achievements from its Integrated Circuit Advanced Process R&D Center. Highlighting the year’s prestige, researcher Ye Tianchun was named a 2026 IEEE Fellow for his pioneering work in advanced device science and its impact on China’s semiconductor industry. Recent breakthroughs include high‑entropy ferroelectric multilayer ceramic capacitors, entanglement‑enhanced nanoscale single‑spin quantum sensors, and reversible shape‑memory two‑dimensional covalent organic frameworks. IME also succeeded in heterogeneously integrating GaN high‑electron‑mobility transistors on 4H‑SiC/diamond composite substrates, paving the way for ultra‑robust power electronics. In parallel, the institute reported a brain‑inspired mechanism that outperforms current AI models, heat‑resistant perovskite‑silicon tandem solar cells, single‑photon light‑beam control, and a photonic chip that splits monochromatic light into three colors. Beyond research, IME offers a full suite of high‑tech platform services—ranging from electron‑beam lithography and atomic‑layer deposition to advanced wafer‑level bonding and cleaning equipment—supporting both academic collaborations and industry partnerships. The center continues to shape China’s microelectronics future while fostering open innovation and talent recruitment.

Read more

Google Turns Old News Stories into AI‑Powered Flash Flood Alerts

Google’s research team has found a clever way to predict flash floods by mining the past. Using its Gemini large‑language model, the team sifted through 5 million news articles from around the globe, pulling out 2.6 million individual flood reports. Each report was tagged with a location and date, creating a massive, geo‑referenced timeline the team calls “Groundsource.” Groundsource gives scientists a real‑world baseline to train a new forecasting model. By feeding the dataset into a Long Short‑Term Memory (LSTM) neural network, the model can combine global weather forecasts with historical flood patterns to calculate the probability of a flash flood in any given area. Google’s product manager Gila Loike says this is the first time the company has used a language model for this kind of environmental data mining. The approach could soon be expanded to other fleeting but dangerous events such as heat waves and mudslides, according to researcher Rothenberg. Industry insiders are taking note. Marshall Moutenot, CEO of Upstream Tech— which already uses deep‑learning to predict river flows for hydropower operators— sees Google’s effort as part of a broader push to build richer, machine‑learning‑ready weather datasets. The Groundsource data and accompanying research were released publicly on Thursday, inviting developers and scientists to build the next generation of AI‑driven weather warnings.

Read more

New AI Tool ‘MatChat 2’ Supercharges Materials Research with Near‑Instant Answers

The Dongguan Institute of Materials has launched MatChat 2, an AI‑driven research assistant that promises to reshape how scientists work with new materials. By using a fresh dynamic‑routing algorithm and smarter knowledge‑retrieval tricks, the system trims average response times down to a few seconds, delivering literature searches and technical guidance almost instantly. MatChat 2’s power comes from a four‑layer collaborative design. A persistent‑context layer keeps conversations flowing for more than ten back‑and‑forth exchanges, while an intent‑classification layer quickly pinpoints what the user really needs and filters out noise. The reasoning‑and‑execution layer balances deep, thoughtful analysis with rapid replies, and a massive academic‑knowledge engine pulls from nearly one million vetted papers to ensure answers are both accurate and relevant. The tool can help at every stage of a materials project— from planning syntheses and running performance tests to drafting scholarly articles and scouting real‑world applications. Most importantly, MatChat 2 dramatically cuts the dreaded “hallucination” problem that plagues large language models. Independent tests show a hallucination rate of just 1.3%, far lower than GPT‑4.1 (5.6%), DeepSeek‑V3 (6.1%) and other leading models. Positioned as a flagship of the institute’s “AI + Science” strategy, MatChat 2 aims to give Guangdong‑Hong Kong‑Macao’s materials‑innovation hub a decisive edge in the global race for next‑generation alloys, ceramics, energy‑storage compounds and more.

Read more