Beijing, China — Researchers from Beijing Institute of Technology and other leading institutions have developed a novel approach for improving Multi-Hop Question Answering (MHQA) tasks, which require models to reason over multiple relevant facts to answer complex questions. The new method, called CausalBridgeQA, integrates causal inference into the MHQA process to address persistent issues of reasoning breakdowns and feature spurious correlations, which are common challenges in current models.

Unlocking Agricultural Potential: New Meta-Analysis Reveals Biochar’s Role in Restoring Salt-Affected Soils

New research provides critical insights into optimizing biochar application for enhancing crop yields and reducing soil salinity in challenged agricultural lands.

A major concern for global food security involves the increasing prevalence of salt-affected soils, which currently encompass an estimated one billion hectares worldwide. Conventional methods for mitigating salt stress can be costly and less effective in the long run. Scientists have focused attention on biochar, a carbon-rich material, as a promising organic soil amendment to improve soil properties and bolster agricultural resilience. A recent meta-analysis, conducted by Baolin Wu, Heng Yang, Siyuan Li, and Jun Tao from Beijing Normal University, delivers the first comprehensive quantitative assessment of biochar's impact on crop productivity and soil salinity in these compromised environments. This analytical effort considered a wide array of experimental conditions, offering a roadmap for tailored biochar applications.

Data is often referred to as the new oil of the digital economy, representing a highly valuable and untapped asset. To fully realize the potential of spatial data, various spatial data marketplace platforms have emerged. The existing spatial data marketplaces primarily focus on recommending each dataset individually. There is a lack of consideration for cases where an individual dataset cannot satisfy the buyer’s needs such that a collection of datasets needs to be acquired.

A comprehensive assessment of Guangdong province's land carbon balance reveals that the highly industrialized region has not yet achieved carbon neutrality, registering a substantial net emission of 925.63 Tg CO₂e in 2021. This significant carbon footprint primarily stems from energy consumption, which accounts for 83.8% of total emissions. Against this backdrop, scientists at Zhejiang University, Guangdong University of Technology, Tsinghua University, and Guangdong Academy of Sciences investigated the potential of biochar technology as a carbon dioxide removal (CDR) strategy, determining its capacity to offset a portion of these emissions. The analysis offers essential guidance for formulating regional emission reduction targets and implementing effective mitigation policies as global temperatures rise.

Researchers from BUPT introduce the RFGDG framework,utilizing RL to dynamically optimize graph generalization in federated settings.

Prompt engineering has emerged as a critical tool to refine AI outputs, but existing techniques are fragmented and lack a cohesive structure. The research team, led by Professor Feng Zhang from Renmin University of China and collaborators from Microsoft AI, Tsinghua University, and the National University of Singapore, published their new research on 15 March 2026 in Frontiers of Computer Science co-published by Higher Education Press and Springer Nature.

The vitality of grassland ecosystems, central to the global carbon cycle and nutrient exchange, is often gauged by their aboveground biomass (AGB). Variations in AGB reflect grassland productivity and overall health. Accurately assessing the diverse factors influencing AGB, particularly distinguishing between influences that play out over decades versus those with immediate effects, has remained an analytical hurdle. Researchers at the Agro-Environmental Protection Institute, Ministry of Agriculture and Rural Affairs, among other institutions, confronted this challenge by developing an advanced statistical framework.