artificial intelligence

AI-driven smart breeding of coarse grain crops: Current status, frontiers and perspectives

Global food security is confronted with mounting pressures arising from continuous population growth, intensifying climate change, and increasing agro-ecological vulnerability. Coarse grain crops, characterized by strong tolerance to drought and poor soil conditions, serve as strategic reserve crops with high nutritional value in the development of diversified food systems. Nevertheless, insufficient research investment and underdeveloped breeding infrastructure have led to the long-term marginalization of many coarse-grain crop resources. Smart breeding, through the integration of biotechnology and artificial intelligence (AI), opens new pathways for breaking through the bottlenecks in coarse grain crops breeding. This review systematically outlined the evolution of smart breeding technologies and discussed the applications of biological big data and AI in the digitization of coarse-grain crop germplasm resources, high-throughput phenomics, genotype-phenotype association analysis, and intelligent decision-making systems. We synthesized representative smart breeding cases in oats, foxtail millet, buckwheat, quinoa, and other coarse grain crops, explored pathways for the industrialization of coarse-grain breeding, and identified persistent challenges—specifically, the shortfall in fundamental research, institutional frictions in the breeding innovation system, and structural constraints in human resources and funding mechanisms and proposed corresponding countermeasures. Finally, we provided an outlook on frontier directions including de novo domestication, genomic design breeding, and digital twins breeding.

Research progress and prospect of soybean smart breeding based on multilayer collaboration of data, algorithms and applications

Soybean [Glycine max (L.) Merr.] is a strategic crop with grain, oilseed, and feed value, yet its improvement has long been constrained by slow yield gains and a narrow genetic base. For China, the high dependence on soybean imports and the low self-sufficiency rate make expanding soybean cultivation and increasing yield per unit area becoming important tasks for ensuring food security, creating an urgent demand for technologies that can improve breeding efficiency. Over the past decade, rapid advances in large-scale sequencing, high-throughput phenotyping, and artificial intelligence algorithms have provided technical conditions for the transformation of breeding approaches. At present, the main constraint on soybean smart breeding is not the insufficiency of any single technology, but the lack of stable connections among data, algorithms, and applications. Whether data can be effectively learned by models, whether prediction results can be translated into breeding decisions such as parent selection, cross screening and target design, and whether new data generated from field validation can be standardizedly fed back, and used to continuously improve models directly affect the operational effectiveness of smart breeding systems. This review summarized research progress in soybean smart breeding from the perspectives of the data layer, algorithm layer, and application layer. It further analyzed the major barriers in data-to-algorithm transfer, algorithm-to-application translation, and application-to-data feedback, and identified key challenges in five aspects: Data, algorithms, applications, platforms and governance. Finally, it discussed the conditions required for soybean breeding to move toward intelligent design breeding. Future efforts should focus on bridging disconnections between layers, making prediction, validation, and data feedback into routine processes, and shifting evaluation criteria from model prediction accuracy to realized genetic gain.

AI-driven innovation and applications in crop breeding technology

The growth rate of global grain production can no longer meet the demands arising from population expansion. Meanwhile, climate change, cultivated land degradation and environmental stress are further exacerbating the vulnerability of the food production system. Ensuring food security is fundamental to maintain the stability and sustainable development of human society. Crop breeding is the pivotal technical means to achieve this goal. Developing new crop varieties with high yield, superior quality, and multiple resistances have become the core pathway to safeguard agricultural sustainable development and global food supply. However, traditional breeding techniques suffer from long cycles, low accuracy and limited efficiency, which cannot satisfy the development requirements of modern seed industry. The rapid iteration of artificial intelligence (AI) technology has injected new intelligent momentum into the innovation of crop breeding, driving the transformation of breeding technology from traditional experience-based breeding to precise, intelligent and efficient breeding. In this review, we summarized the development of crop breeding, and focused on the innovative breakthroughs of genomic selection, precision genome editing, protein design and high-throughput phenotyping driven by AI. We also further elaborated the intelligent driving effects exerted by these technologies on key breeding links involving germplasm mining, gene function analysis, directional trait improvement, intelligent phenotypic assessment and intelligent factory breeding. Finally, we discussed the challenges and future developmental prospects of AI deployment in crop breeding, aiming to provide a reference for the innovation and industrial application of intelligent breeding technologies.

AI explained: a guide for non-technical readers

AI Explained: A Guide for Non-Technical Readers builds understanding of artificial intelligence from first principles rather than diving in at the top. Written by experienced policy and technical experts, the book walks through rules and logic-based approaches, statistical methods, neural networks, machine learning, and generative models in accessible, structured terms. Each major section concludes with a dedicated use-cases chapter grounding abstract concepts in practical scenarios drawn from healthcare, law, and business. Rather than teaching readers how to build or deploy AI, the book answers a more fundamental question: how do these systems achieve the outcomes they produce? Coverage of AI policy, ethics, and societal impact rounds out the treatment, informed directly by the authors' advisory roles with governments and international bodies.

AI analysis of elastin staining sections reveals the biological implications of cancer-related fibrosis in small-sized lung adenocarcinomas

Objective: Pathological invasion influences patient prognosis in lung adenocarcinoma (LUAD); however, diagnosis is often associated with high interobserver variability because non-lepidic adenocarcinoma (NLA) and cancer-related fibrosis (CRF) are intricately mixed, and CRF contains both invasive and non-invasive components. In this study, to enable quantitative and reproducible diagnosis of areas showing high intratumoral heterogeneity, we proposed an artificial intelligence (AI)-based analysis distinguishing between NLA and CRF. This approach could reduce interobserver variability in prognostic prediction and help examine the different biological meanings of NLA and CRF. Methods: To clarify the physiological structure of lung parenchyma, we used elastin staining specimens. CRF and NLA were separately annotated in the first cohort (n = 35) and used for supervised learning. The AI then analyzed whole non-lepidic areas in the second cohort (n = 188); we then examined the relationship between clinicopathological features and the AI analysis. Results: For the first cohort, the accuracy was 89.2% on average. For the second cohort, groups with high CRF ratios (>50%) in the AI analysis showed a statistical correlation with types B-C in Noguchi's classification (p < 0.001), grades 1-2 in the WHO grading system (p = 0.018), and a higher 5-year disease-free survival (DFS) rate (85.7% vs. 69.4%, p = 0.006). In groups with a low CRF ratio (n = 111), a central distribution of CRF was associated with a lower DFS rate (41.2% vs. 74.5%, p = 0.001). Conclusion: The AI engine for distinguishing between NLA and CRF enabled prognostic prediction of LUAD with reduced interobserver variability. It may also be useful for examining the biological significance of NLA and CRF.

How maths changes the world: mathematical science impact case studies drawn from REF 2021

The mathematical sciences are an essential—albeit often invisible—part of the UK's national infrastructure, woven through almost every sector of the economy and most of the systems and services on which society depends. They are fundamental to our resilience, quality of life, and prosperity: recent analysis by AcadMathSci found that people working in mathematical science occupations directly contributed nearly half a trillion pounds to the UK economy in 2024, equivalent to about 19% of total GDP. But what does this contribution mean in practice? One does not have to look far for examples. The Research Excellence Framework (REF) is the UK's system for assessing the quality and impact of research carried out in higher education institutions. Its most recent exercise, REF 2021, included many detailed impact case studies showing how university research has delivered tangible benefits far beyond academia: from improving public services and protecting lives to supporting new technologies, strengthening businesses, and informing public policy. This report distils a subset of the 176 publicly available REF 2021 impact case studies associated with the mathematical sciences$^1$ into concise, accessible summaries. The case studies reach across the economy and society, showing how mathematical ideas developed in one setting can provide the methods, evidence, and insight needed to solve important problems elsewhere. Together, they illustrate both the breadth of the UK's mathematical research and the variety of ways in which it is changing the world. This collection is necessarily a selective and time-bound snapshot, rather than a comprehensive account, of the mathematical sciences' leading-edge impact. For example, REF 2021 captured research and impact developed before the recent rapid advances in artificial intelligence (AI) capabilities: AI-related case studies were thus far less prominent than they are likely to be in future exercises. Work connected to national security is also substantially under-represented: a REF 2021 overview report noted that many highly confidential case studies related to national security (i.e., not among the 176 mathematical science case studies$^1$ in the public domain) were based on "mathematical research of the very highest calibre." Despite such constraints, these case studies point consistently to the same conclusion. When we invest in the mathematical sciences—in education, research, and people—we strengthen the UK’s capacity to innovate, respond to uncertainty, and improve lives. That investment helps build a more productive, resilient, and prosperous future, while generating knowledge and tools with benefits reaching far beyond the problems they were first developed to solve. Maximising the UK-wide impact of the mathematical sciences, and sustaining that impact into the future, requires national capacity to identify, connect, and utilise mathematical expertise. AcadMathSci is working to build that national capacity by bringing together education, academia, business, industry and government; translating evidence into forms useful for decision-makers, policymakers, and industry; and helping ensure that the mathematical sciences are recognised and supported as a strategic national asset. With sustained support, AcadMathSci will be able to advance its wider work to strengthen the mathematical sciences across the UK. Repeating and extending this exercise following REF 2029 would form one part of that work: capturing a new generation of impact case studies, strengthening the evidence base for support for the mathematical sciences, and revealing many further ways in which the mathematical sciences are transforming lives, society, and the economy.

How maths changes the world: mathematical science impact case studies drawn from REF 2021

The mathematical sciences are an essential—albeit often invisible—part of the UK's national infrastructure, woven through almost every sector of the economy and most of the systems and services on which society depends. They are fundamental to our resilience, quality of life, and prosperity: recent analysis by AcadMathSci found that people working in mathematical science occupations directly contributed nearly half a trillion pounds to the UK economy in 2024, equivalent to about 19% of total GDP. But what does this contribution mean in practice? One does not have to look far for examples. The Research Excellence Framework (REF) is the UK's system for assessing the quality and impact of research carried out in higher education institutions. Its most recent exercise, REF 2021, included many detailed impact case studies showing how university research has delivered tangible benefits far beyond academia: from improving public services and protecting lives to supporting new technologies, strengthening businesses, and informing public policy. This report distils a subset of the 176 publicly available REF 2021 impact case studies associated with the mathematical sciences$^1$ into concise, accessible summaries. The case studies reach across the economy and society, showing how mathematical ideas developed in one setting can provide the methods, evidence, and insight needed to solve important problems elsewhere. Together, they illustrate both the breadth of the UK's mathematical research and the variety of ways in which it is changing the world. This collection is necessarily a selective and time-bound snapshot, rather than a comprehensive account, of the mathematical sciences' leading-edge impact. For example, REF 2021 captured research and impact developed before the recent rapid advances in artificial intelligence (AI) capabilities: AI-related case studies were thus far less prominent than they are likely to be in future exercises. Work connected to national security is also substantially under-represented: a REF 2021 overview report noted that many highly confidential case studies related to national security (i.e., not among the 176 mathematical science case studies$^1$ in the public domain) were based on "mathematical research of the very highest calibre." Despite such constraints, these case studies point consistently to the same conclusion. When we invest in the mathematical sciences—in education, research, and people—we strengthen the UK’s capacity to innovate, respond to uncertainty, and improve lives. That investment helps build a more productive, resilient, and prosperous future, while generating knowledge and tools with benefits reaching far beyond the problems they were first developed to solve. Maximising the UK-wide impact of the mathematical sciences, and sustaining that impact into the future, requires national capacity to identify, connect, and utilise mathematical expertise. AcadMathSci is working to build that national capacity by bringing together education, academia, business, industry and government; translating evidence into forms useful for decision-makers, policymakers, and industry; and helping ensure that the mathematical sciences are recognised and supported as a strategic national asset. With sustained support, AcadMathSci will be able to advance its wider work to strengthen the mathematical sciences across the UK. Repeating and extending this exercise following REF 2029 would form one part of that work: capturing a new generation of impact case studies, strengthening the evidence base for support for the mathematical sciences, and revealing many further ways in which the mathematical sciences are transforming lives, society, and the economy.

Personalizing Assignments in the AI Era: The Role of Engaged Pedagogy

Generative artificial intelligence (Gen AI) has taken the academic world by storm. This G.I.F.T.S. paper argues that personalizing academic assignments may curb students’ tendency to copy and paste content from Gen AI outputs. I demonstrate how I personalized a written assignment in one of my courses. A total of 81 paper grades were analyzed, and no significant difference was found between students who engaged in personalizing their learning and those who did not. Final course grades showed the same result. Although no statistically significant difference was detected, personalizing assignments remains a valuable way to engage students in the Gen AI era.

Progress in mantle anisotropy research based on XKS phases

XKS shear-wave splitting has been widely applied to investigations of anisotropy in both the upper mantle and the lowermost mantle D" layer, and represents a key tool for probing mantle deformation and dynamics. The first part of this paper reviews the basic principles and major approaches of XKS splitting analysis, and summarizes its applications to upper-mantle anisotropy beneath the Chinese mainland, where strong regional variations in deformation are observed, primarily controlled by the subduction of the Indian plate and the western Pacific plate. We further discuss methods, applications, and challenges in using XKS phases to probe D" anisotropy, highlighting that complex and strong anisotropic structures are commonly observed beneath remnant slab regions and large low-shear-velocity provinces, while effective separation of upper-mantle contributions and mitigation of wavefield scattering and interference remain major difficulties. Over the past decade, with the rapid growth of seismic datasets and advances in inversion techniques, XKS-based three-dimensional anisotropic imaging of the upper mantle has become increasingly mature, significantly improving depth resolution of anisotropic structures. The second part of this paper introduces the theoretical framework and development of this imaging approach and summarizes recent results. This method is capable of delineating three-dimensional upper-mantle anisotropy and effectively identifying deep structures such as slab geometries, mantle upwellings, and layered anisotropy, although the imaging quality strongly depends on data coverage and observation density. Finally, we discuss future perspectives of XKS studies in mantle anisotropy research. The field is evolving from traditional parameter-based analyses toward full three-dimensional imaging. However, three-dimensional anisotropic imaging of both the upper mantle and the D" layer beneath the Chinese mainland remains limited. Future efforts should integrate multiple anisotropic constraints, promote multi-layer mantle coupling studies, and incorporate artificial intelligence techniques to improve the accuracy of shear-wave splitting analysis, thereby providing a foundation for constructing high-resolution, high-precision three-dimensional anisotropic models of the mantle beneath China.

Artificial intelligence for detection, grading, and prognostication in prostate cancer pathology: A scoping review.

Artificial intelligence (AI) has been transforming many aspects of medical care. In prostate cancer, ongoing progress in AI has improved research and patient care. Recent advances in machine learning and deep learning have produced tools that help diagnose cancer, assess risk, and predict outcomes. In screening, AI-based risk calculators improve detection and help avoid unnecessary biopsies. Deep learning algorithms, particularly convolutional neural networks, have demonstrated expert-level performance in pathology, identifying malignancy and assigning Gleason grades with high accuracy. These tools also streamline workflow, flagging challenging cases for review and quantifying prognostic markers, such as Ki-67 and cribriform patterns. In addition, AI-based models can predict molecular alterations, microsatellite instability, and lymph node metastasis directly from histology images, providing cost-effective alternatives to traditional assays. The development of multimodal models integrates digital pathology and clinical parameters, enabling personalized treatment recommendations and improved outcome prediction. Natural language processing and large language models further expand AI's potential, facilitating information extraction from clinical notes and enhancing patient education. Despite these advances, most studies remain retrospective with heterogeneous endpoints. Performance often drops when models are tested at new sites because of differences in patient populations and slide preparation. Access to large, well-annotated datasets is limited, and technical variation hampers reproducibility. To move toward clinical use, the field needs prospective, multicenter validation, preanalytical and analytical standardization, and clear reporting of failure modes and human oversight. Emerging approaches, including self-supervised pretraining, transformer-based image models, and language-vision systems, are likely to improve generalization and support more personalized care.

Integration of functional magnetic resonance imaging and artificial intelligence in Alzheimer’s disease

Linguistic Justice and Artificial Intelligence