Nature Machine Intelligence最新文献

{"title":"A manifesto for Sustainability Robotics","authors":"S. Song, B. Mazzolai, M. Kovač","doi":"10.1038/s42256-026-01260-6","DOIUrl":"10.1038/s42256-026-01260-6","url":null,"abstract":"Sustainability spans environmental, societal and economic challenges, from climate change to healthcare and education. Robotics holds substantial promise for addressing these issues, yet current developments remain fragmented and lack a unifying framework. This fragmentation risks unintended consequences, including unequal access to technology and missed opportunities for broader impact. Here we advocate for a new discipline, Sustainability Robotics, structured around three guiding principles: robots should be minimally invasive, reducing disruption to ecosystems and socio-economic systems; universally accessible, extending benefits to underserved communities and extreme environments; and symbiotic, generating mutually beneficial outcomes for humans and nature. We define two complementary dimensions. The first, sustainable robot design, focuses on minimizing environmental impact through materials, energy and manufacturing. The second, robotic solutions for sustainability, leverages robotics to address environmental, social and economic challenges. By integrating perspectives from science, engineering, economics, ethics and policy, Sustainability Robotics provides a foundation for coordinated research, education and innovation. This framework aims to align robotics development with global sustainability goals, enabling more equitable and effective technological impact. Song et al. propose Sustainability Robotics as a new discipline to overcome fragmentation and enhance societal and environmental impact. They define three guiding principles, alongside two dimensions spanning sustainable design and robotics for sustainability.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 7","pages":"1038-1044"},"PeriodicalIF":29.8,"publicationDate":"2026-07-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148432200","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Towards shared embodied intelligence in humanoid robots through optimization, development and testing of the human-aware ergoCub robot","authors":"Carlotta Sartore, Mohamed Elobaid, Lorenzo Rapetti, Giulio Romualdi, Stefano Dafarra, Nicola A. Piga, Ines Sorrentino, Paolo Maria Viceconte, Silvio Traversaro, Ugo Pattacini, Luca Fiorio, Francesco Draicchio, Giovanna Tranfo, Lorenzo Natale, Marco Maggiali, Daniele Pucci","doi":"10.1038/s42256-026-01272-2","DOIUrl":"10.1038/s42256-026-01272-2","url":null,"abstract":"Collaboration is central to human behaviour, enabling tasks beyond individual capability. This ability arises from coordinating actions through internal representations of others, a concept known as shared intelligence. In addition, humans are characterized by physical bodies and cognitive abilities that are optimized in response to their environment, a phenomenon referred to as embodied cognition. Designing humanoid robots that collaborate safely and effectively with people requires unifying these principles. Here we propose an architecture that integrates shared intelligence and embodied cognition to enable robots to physically collaborate with humans, where robot hardware and control are optimized for human metrics, using representations of the human body and motion intelligence. The ultimate goal is to achieve a form of shared embodied intelligence. Specifically, our architecture optimizes robot hardware and physical intelligence parameters with respect to human ergonomic metrics. This is accomplished by modelling human–robot interaction as a function of hardware configurations and embedding human models into the robot’s physical intelligence. As a concrete implementation, we present the humanoid robot ergoCub, whose morphology and control have been optimized for collaborative tasks with humans. Our approach provides a framework for designing humanoid robots that prioritize human ergonomics at both the hardware and physical intelligence levels, with applications in industrial and assistive robotics. Sartore et al. present ergoCub, a humanoid robot that prioritizes human safety at hardware and motion levels. Using a shared embodied intelligence framework, design and control are jointly optimized with human-related metrics such as back stress alongside locomotion objectives, reducing spinal load and improving walking robustness.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 8","pages":"1221-1237"},"PeriodicalIF":29.8,"publicationDate":"2026-07-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.comhttps://www.nature.com/articles/s42256-026-01272-2.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148432201","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Four questions for AI-ready biological data","authors":"Amarda Shehu, Ruth Nussinov","doi":"10.1038/s42256-026-01270-4","DOIUrl":"10.1038/s42256-026-01270-4","url":null,"abstract":"The push to make biological data ‘AI-ready’ is accelerating worldwide. But to understand what AI-ready means for biological data, there are several questions to answer.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 7","pages":"1017-1019"},"PeriodicalIF":29.8,"publicationDate":"2026-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148432236","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Guiding generative models to uncover diverse and novel crystals via reinforcement learning","authors":"Hyunsoo Park, Aron Walsh","doi":"10.1038/s42256-026-01262-4","DOIUrl":"10.1038/s42256-026-01262-4","url":null,"abstract":"Discovering functional crystalline materials entails navigating an immense combinatorial design space. Although recent advances in generative artificial intelligence have enabled the sampling of chemically plausible compositions and structures, a fundamental challenge remains: the objective misalignment between the likelihood-based sampling in generative modelling and the targeted focus on underexplored regions where novel compounds reside. Here we introduce a reinforcement learning framework that guides latent denoising diffusion models in finding diverse and novel, yet thermodynamically viable, crystalline compounds. Our approach integrates group-relative policy optimization with verifiable, multi-objective rewards that jointly balance creativity, stability and diversity. Beyond de novo generation, we demonstrate enhanced property-guided design that preserves chemical validity while targeting desired functional properties. This approach establishes a modular foundation for controllable AI-driven inverse design that addresses the novelty–validity trade-off across the scientific discovery applications of generative models. Park and Walsh introduce a reinforcement learning framework that could accelerate the discovery of new, thermodynamically stable and diverse crystalline materials with desired properties.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 7","pages":"1087-1099"},"PeriodicalIF":29.8,"publicationDate":"2026-07-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.comhttps://www.nature.com/articles/s42256-026-01262-4.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148409205","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Principled approaches for extending neural architectures to function spaces for operator learning","authors":"Julius Berner, Miguel Liu-Schiaffini, Jean Kossaifi, Valentin Duruisseaux, Boris Bonev, Kamyar Azizzadenesheli, Anima Anandkumar","doi":"10.1038/s42256-026-01267-z","DOIUrl":"10.1038/s42256-026-01267-z","url":null,"abstract":"Deep learning has achieved remarkable success in computer vision and natural language processing, where tasks are commonly formulated as mappings between finite-dimensional representations. Many scientific problems, however, including those governed by partial differential equations, are naturally posed on infinite-dimensional function spaces. This mismatch has limited conventional neural networks from achieving comparable success in scientific applications. Here we identify and distil key principles for constructing practical neural architectures for mappings between function spaces. Neural operators provide a principled extension of neural networks to such settings, offering a path towards bringing deep learning’s transformative impact to science. Because deep learning’s success has relied heavily on architectural refinements, extending these advances to neural operators allows operator learning to benefit from refined designs. Guided by the principles we outline, we propose a recipe for converting popular neural architectures into neural operators with minimal modifications. We also discuss practical steps for making these models effective. This perspective offers a systematic bridge between finite-dimensional network design and operator learning for scientific applications. Berner et al. show how to adapt popular neural networks into discretization-agnostic neural operators that learn from continuous scientific data, enabling scientific simulations that generalize more reliably across resolutions.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 7","pages":"1173-1181"},"PeriodicalIF":29.8,"publicationDate":"2026-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.comhttps://www.nature.com/articles/s42256-026-01267-z.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148385804","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Reshaping biomolecular structure prediction through strategic conformational exploration with HelixFold-S1","authors":"Lihang Liu, Yang Liu, Xianbin Ye, Shanzhuo Zhang, Yuxin Li, Kunrui Zhu, Yang Xue, Jingbo Zhou, Xiaonan Zhang, Xiaomin Fang","doi":"10.1038/s42256-026-01264-2","DOIUrl":"10.1038/s42256-026-01264-2","url":null,"abstract":"Generating large ensembles of candidate conformations is standard for improving biomolecular structure prediction. Yet aimless sampling is inefficient and costly, producing many redundant conformations with limited diversity, particularly for complex multimeric assemblies. Here we present HelixFold-S1, a guided planning approach specifically designed to enhance the structural prediction of biomolecular complexes by strategically targeting the most informative regions of conformational space to produce accurate conformations. For each complex, predicted interchain contact probabilities serve as a blueprint of the conformational space, guiding computational effort towards higher-probability, low-redundancy contacts that constrain structure generation. Across diverse biomolecular complex benchmarks, HelixFold-S1 achieves markedly higher structural accuracy than traditional unguided methods while reducing sampling requirements by an order of magnitude. Predicted contact probabilities also provide a rough indicator of prediction difficulty and sampling utility. These results demonstrate that guided planning reshapes conformational exploration and enables more efficient and accurate structural inference. Liu and colleagues introduce HelixFold-S1, a guided sampling strategy for biomolecular complex structure prediction that targets high-probability interaction regions. The method achieves higher accuracy than traditional unguided methods while reducing computational costs.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 7","pages":"1115-1126"},"PeriodicalIF":29.8,"publicationDate":"2026-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148385806","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Empowering biomedical evidence exploration and synthesis with deep knowledge graph research","authors":"Zifeng Wang, Zheng Chen, Ziwei Yang, Xuan Wang, Qiao Jin, Yifan Peng, Zhiyong Lu, Jimeng Sun","doi":"10.1038/s42256-026-01266-0","DOIUrl":"10.1038/s42256-026-01266-0","url":null,"abstract":"Compared with generic artificial intelligence agents, deep research agents perform longer-horizon reasoning and deeper literature exploration to investigate complex questions. Here we present DeepEvidence, a deep research agent for evidence exploration and synthesis across heterogeneous biomedical knowledge sources. DeepEvidence advances deep research through coordinated multi-agent collaboration combining breadth-first and depth-first research strategies to search, explore and aggregate evidence from multiple biomedical knowledge bases and literature. It also incrementally constructs an evidence graph of key entities and observations to support transparent tracking, attribution and validation of the research process. DeepEvidence substantially outperforms generic artificial intelligence agents across four open benchmarks. We further establish seven benchmark tasks spanning major stages of biomedical discovery, including drug discovery, preclinical experimentation, clinical trial development and evidence-based medicine. DeepEvidence demonstrates substantial improvements in systematic evidence exploration and synthesis. These results highlight the potential of deep research agents to accelerate biomedical discovery and translational research. Wang et al. develop DeepEvidence, a biomedical deep research agent for exploring and synthesizing evidence across various knowledge sources to support drug discovery, clinical trials and evidence-based medicine.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 7","pages":"1142-1156"},"PeriodicalIF":29.8,"publicationDate":"2026-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148385805","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An agentic artificially intelligent X-ray scientist","authors":"Zhantao Chen, Alexander N. Petsch, Aidan J. Israelski, Rajan Plumley, Lingjia Shen, Cong Wang, Cheng Peng, Yuan Ni, Arun Bansil, Sugata Chowdhury, Mingda Li, Jana B. Thayer, Vivek Thampy, Joshua J. Turner","doi":"10.1038/s42256-026-01261-5","DOIUrl":"10.1038/s42256-026-01261-5","url":null,"abstract":"Executing experimental tasks in both normal research laboratories and large-scale scientific facilities often requires extensive human supervision and remains a key challenge on the path to fully autonomous, artificial intelligence (AI)-driven science. Here we demonstrate a large language model-driven agent that autonomously performs X-ray sample alignment on a synchrotron beamline by planning actions, executing instrumental commands, interpreting observations and iterating towards experimental goals. Based on existing large language models with structured tool-use via the model context protocol, our AI X-ray scientist was guided and tested using an in-house-built virtual experimental setup that mirrors a six-circle diffractometer at an operational synchrotron beamline. The agentic workflow developed in the virtual environment was directly deployed on a real beamline, where it correctly identified reference reflections and determined the orientation matrix, an essential first step in any type of single-crystal scattering experiment. Our AI X-ray scientist responded effectively to unexpected experimental conditions, demonstrating adaptive problem-solving and readiness for addressing practical experimental situations. Our study provides a step towards autonomous operation across diverse experimental environments at large-scale scattering facilities. Chen et al. demonstrate an AI X-ray scientist that autonomously aligns single crystals at a real synchrotron beamline, showing how large language models can enable adaptive closed-loop experimentation at large-scale scientific facilities.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 7","pages":"1075-1086"},"PeriodicalIF":29.8,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.comhttps://www.nature.com/articles/s42256-026-01261-5.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148385803","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Data-driven surrogates of rational design enable antimicrobial peptide optimization","authors":"Goran Mauša, Daniela Kalafatovic","doi":"10.1038/s42256-026-01258-0","DOIUrl":"10.1038/s42256-026-01258-0","url":null,"abstract":"Rising pathogen drug resistance makes next-generation antimicrobial peptides a global priority. Generative AI accelerates discovery by rapidly proposing new peptides with high therapeutic potential. The key question is no longer whether broad data-driven exploration is possible, but whether it can refine biologically complex activity scaffolds.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 7","pages":"1020-1022"},"PeriodicalIF":29.8,"publicationDate":"2026-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148617412","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Solutions, challenges and rising tensions in AI and mathematics","authors":"","doi":"10.1038/s42256-026-01269-x","DOIUrl":"10.1038/s42256-026-01269-x","url":null,"abstract":"Recent breakthroughs in mathematical research show that AI is transforming the field at a remarkable pace. In an open letter published this month, an international group of mathematicians argue that the field needs to remain a human endeavour.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 6","pages":"857-857"},"PeriodicalIF":29.8,"publicationDate":"2026-06-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.comhttps://www.nature.com/articles/s42256-026-01269-x.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148300200","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}