A research team from Merck (China) has proposed Jigsaw-LightRAG, a novel method that improves how artificial intelligence systems manage and update knowledge graphs.  It decomposes a global knowledge graph (KG) into per-document subgraphs and leverages document lifecycle states (New, Modified, Persistent, Deleted) so that LLM-based extraction is invoked only for changed documents. The updated global KG is then reconstructed via code-level aggregation and strict string-matching deduplication without additional LLM token consumption. Experiments across public benchmarks (LongBench, PubMedQA, and QASPER) demonstrate that in localized corpus changing scenarios, LLM token consumption drops by orders of magnitude (with DELETE operations requiring 0 tokens), while KG structural integrity remains stable and question-answering performance stays on par with full-rebuild baselines.

A research team led by Professor Taesung Kim from the School of Mechanical Engineering has developed a technology that precisely controls the internal structure of semiconductors using heat, much like stamping out “bungeoppang” (fish-shaped pastry) in a mold. The team reported that this approach improves the performance of next-generation artificial intelligence (AI) hardware. With this technology, complex AI computations can be processed more quickly using significantly less electricity than before.

Fighting fires could be done remotely without the need to place firefighting crews directly in potentially dangerous situations by using collaborative teams of artificial intelligence-powered robots with extinguishing equipment on board, with an initial soft trial of the technology proving successful.

Scientists have developed an AI-powered multi-agent system, named eNRRCrew, to tackle the challenge of producing sustainable ammonia. By automatically mining and analyzing over 2,300 scientific papers, this AI "crew" built a massive database and used it to identify key factors for designing efficient catalysts. The system can predict catalyst performance, recommend novel materials, and has already proposed a promising candidate that was validated by computational simulations and experiments. This AI-driven approach offers a new paradigm for accelerating scientific discovery in catalysis and beyond.

As an important quantum communication protocol, quantum teleportation has broad applications in quantum information science and technology. Toward the application, it is essential to enhance the ability of quantum teleportation by teleporting multiple quantum states simultaneously. Recently, the group led by Xiaolong Su at Shanxi University presents controllable deterministic quantum teleportation of multiple sideband qumodes simultaneously with the assistance of continuous-variable quantum entanglement. They show that the number of teleportable sideband qumodes in continuous-variable quantum teleportation are controllable by tuning the phase of the classical channel. The work presents a novel method for quantum teleportation of multiple quantum states simultaneously and take a crucial step in enhancing the teleporting ability of quantum teleportation.

Researchers at Beijing Institute of Technology have experimentally demonstrated quantized soliton pumping governed by higher-dimensional Chern invariants. Using a time-modulated lattice and its corresponding topolectrical-circuit platform, they observe two-dimensional soliton pumping whose per-cycle displacement is determined by the first and second Chern numbers. Published in National Science Review, this work reveals that, by tuning the nonlinear strength and the linear band structure, the pumping can transition among integer, fractional, trapped, and anisotropic integer–fractional regimes.

Chinese researchers have developed China's first compact high-temperature superconducting magnetoplasmadynamic thruster, achieving a revolutionary reduction in power consumption from 285 kW to under 1 kW and weight from 220 kg to 60 kg. Published in National Science Review, the breakthrough utilizes YBCO superconducting material operating at liquid nitrogen temperatures (-196°C) to replace traditional copper coils. The thruster demonstrates an exceptional specific impulse of 3,265 seconds at 12 kW input power—more than ten times higher than conventional chemical propulsion. The team also established a comprehensive analytical magnetohydrodynamic model that accurately predicts performance parameters. This advancement solves the critical propulsion bottleneck for small satellites, enabling lightweight, cost-effective spacecraft with dramatically reduced fuel requirements for deep space missions.

A new study introduces China's Glacial Lake Management System (GLMS), an integrated approach that combines monitoring, engineering, and community engagement to reduce the impacts of glacial lake outburst floods (GLOFs) in the Himalaya. Research shows that while GLOF frequency has not significantly increased, downstream damage has risen due to growing exposure. The GLMS, implemented since 2019, has helped mitigate losses in China. Modeling indicates that without intervention, future GLOF intensity could rise by over 27%, but GLMS measures could reduce flood intensity by nearly one-quarter. The study proposes a scalable framework for regional cooperation to protect vulnerable communities across the Himalaya.

A large-scale synchronized observation across 37 cities in North China Plain reveals a critical transition in urban ozone formation mechanisms, from predominantly VOC-limited regime to VOC–NO_x co-limited regime. This finding indicates that China’s ozone control efforts have passed through the most challenging phase, in which NO_x reductions tended to exacerbate ozone pollution. Further reductions in both NO_x and VOC emissions are now expected to be broadly effective, offering greater flexibility, feasibility, and optimism for future ozone mitigation in China.