Nature Machine Intelligence最新文献_第2页

Regional climate risk assessment from climate models using probabilistic machine learning 基于概率机器学习的气候模型的区域气候风险评估

IF 29.8 1区计算机科学

Nature Machine Intelligence Pub Date : 2026-09-28 DOI: 10.1038/s42256-026-01308-7

Zhong Yi Wan, Ignacio Lopez-Gomez, Robert Carver, Tapio Schneider, John Anderson, Fei Sha, Leonardo Zepeda-Núñez

{"title":"Regional climate risk assessment from climate models using probabilistic machine learning","authors":"Zhong Yi Wan, Ignacio Lopez-Gomez, Robert Carver, Tapio Schneider, John Anderson, Fei Sha, Leonardo Zepeda-Núñez","doi":"10.1038/s42256-026-01308-7","DOIUrl":"10.1038/s42256-026-01308-7","url":null,"abstract":"Effective climate risk assessment is hindered by the resolution gap between coarse global climate models and the fine-scale information needed for regional decision making. Existing downscaling methods struggle to bridge this gap: physics-based methods are too computationally expensive for large ensemble sampling of extreme events, while traditional statistical methods fail to capture multivariate spatiotemporal dependencies crucial to compound risk estimation. Furthermore, current machine learning approaches rely on temporally aligned training pairs, which are unavailable for free-running climate projections. Here we introduce GenFocal, an AI framework that generates statistically accurate, fine-scale weather from coarse climate projections without requiring paired simulated and observed events during training. GenFocal synthesizes complex and long-lived hazards, such as heatwaves and tropical cyclones, even when they are not well represented in the coarse climate projections. It also samples high-impact, rare events more accurately than leading methods. By translating large-scale climate projections into actionable localized information, GenFocal provides a powerful paradigm to improve climate adaptation and resilience strategies. A generative AI framework called GenFocal is introduced for climate downscaling, producing realistic fine-scale weather from coarse projections and improving regional risk estimates of compound extremes such as heatwaves and tropical cyclones.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 10","pages":"1560-1568"},"PeriodicalIF":29.8,"publicationDate":"2026-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.comhttps://www.nature.com/articles/s42256-026-01308-7.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149087643","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}

引用次数: 0

Task-structured modularity emerges in artificial networks and aligns with brain architecture 任务结构模块化出现在人工网络中，并与大脑结构保持一致

IF 29.8 1区计算机科学

Nature Machine Intelligence Pub Date : 2026-09-28 DOI: 10.1038/s42256-026-01306-9

Yuhang Wu, Shikuang Deng, Kangrui Du, Marcelo G. Mattar, Yifan Wu, Dani S. Bassett, Huajin Tang, Gang Pan, Shi Gu

{"title":"Task-structured modularity emerges in artificial networks and aligns with brain architecture","authors":"Yuhang Wu, Shikuang Deng, Kangrui Du, Marcelo G. Mattar, Yifan Wu, Dani S. Bassett, Huajin Tang, Gang Pan, Shi Gu","doi":"10.1038/s42256-026-01306-9","DOIUrl":"10.1038/s42256-026-01306-9","url":null,"abstract":"Understanding how neural systems develop modular organization is fundamental to both neuroscience and artificial intelligence. Modular architectures are thought to be critical for adaptability and complex cognitive performance by enriching the brain’s ability to integrate and process information. While prior work posits that modularization emerges from physical constraints, such as minimizing metabolic wiring costs, these spatial models alone do not fully explain the functional organization of brain networks. A critical, unanswered question is how the functional demands of learning complex tasks computationally influence network topology. Here we demonstrate, using recurrent neural networks trained on cognitive tasks, how functional demands shape structural organization. We find that multitask learning paradigms enhance network modularity compared with single-task training, especially when task load strains the network’s capacity. Networks trained with incremental multitask learning, in particular, develop the highest degree of modularity and maintain superior performance. Furthermore, these task-induced networks exhibit structural properties that more closely resemble biological brain networks than models based solely on spatial constraints. These results provide a controlled computational account showing that functional demands can induce modular organization in artificial recurrent networks under capacity constraints, suggesting modularization as an adaptive response to the sequential introduction of complex tasks. Wu et al. show that neural networks learning multiple tasks organize into specialized modules, particularly under capacity constraints. Incremental multitask learning strengthens this modularity and produces architectures resembling brain networks.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 10","pages":"1548-1559"},"PeriodicalIF":29.8,"publicationDate":"2026-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149087646","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}

引用次数: 0

Guiding large language models to predict edit sequences for molecular synthesizability optimization 指导大型语言模型预测编辑序列，优化分子合成能力

IF 29.8 1区计算机科学

Nature Machine Intelligence Pub Date : 2026-09-23 DOI: 10.1038/s42256-026-01304-x

Junren Li, Luhua Lai

{"title":"Guiding large language models to predict edit sequences for molecular synthesizability optimization","authors":"Junren Li, Luhua Lai","doi":"10.1038/s42256-026-01304-x","DOIUrl":"10.1038/s42256-026-01304-x","url":null,"abstract":"Generative artificial intelligence has revolutionized the exploration of chemical space, yet a critical bottleneck remains that a substantial fraction of generated molecules is synthetically inaccessible. Current solutions, such as post hoc filtering or projection-based methods, often compromise structural novelty or disrupt key pharmacophores by forcing molecules into predefined synthetic templates. Here we introduce SynCraft, a reasoning-based framework that reframes synthesizability optimization not as a sequence translation task but as a precise structural editing problem. Leveraging the emergent reasoning capabilities of large language models, SynCraft navigates the ‘synthesis cliff’ where minimal structural modifications yield substantial gains in synthetic feasibility. By predicting executable sequences of atom-level edits rather than generating SMILES strings directly, SynCraft circumvents the syntactic fragility of LLMs while harnessing their chemical intuition. Extensive benchmarks demonstrate that SynCraft outperforms state-of-the-art baselines in generating synthesizable analogues with high structural fidelity, with comparable performance across both proprietary and open-weight backends. Furthermore, through interaction-aware prompting, SynCraft successfully replicates expert medicinal chemistry intuition in editing PLK1 inhibitors and rescuing high-scoring RIPK1 candidates previously discarded in the molecular generation literature. For application convenience, we package the framework as an agent skill and demonstrate an end-to-end rescue workflow on candidate molecules generated for the SARS-CoV-2 main protease as an example. Computational drug design often generates molecules that cannot be synthesized. Li and Lai show how large language models can predict precise structural edits to fix these molecules, outperforming traditional methods and preserving key features.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 10","pages":"1537-1547"},"PeriodicalIF":29.8,"publicationDate":"2026-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149021378","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}

引用次数: 0

Cybernetics, interoception, and the art of embodiment 控制论，内感受，和化身的艺术

IF 29.8 1区计算机科学

Nature Machine Intelligence Pub Date : 2026-09-18 DOI: 10.1038/s42256-026-01312-x

引用次数: 0

Publisher Correction: Quantum neural operators with implicit quadratic frame and expressivity advantages 发布者更正：具有隐式二次框架和表达性优势的量子神经算子

IF 29.8 1区计算机科学

Nature Machine Intelligence Pub Date : 2026-09-14 DOI: 10.1038/s42256-026-01314-9

Ruocheng Wang, Xiaoqiu Zhong, Zhuo Xia, Junchi Yan

引用次数: 0

A collaborative agent with two lightweight synergistic models for autonomous crystal materials research 自主晶体材料研究中两个轻量级协同模型的协同代理

IF 29.8 1区计算机科学

Nature Machine Intelligence Pub Date : 2026-09-10 DOI: 10.1038/s42256-026-01298-6

Tongyu Shi, Yutang Li, Zhanyuan Li, Qian Liu, Jie Zhou, Wenhe Xu, Yang Li, Dawei Dai, Rui He, Wenhua Zhou, Jiahong Wang, Xue-Feng Yu

{"title":"A collaborative agent with two lightweight synergistic models for autonomous crystal materials research","authors":"Tongyu Shi, Yutang Li, Zhanyuan Li, Qian Liu, Jie Zhou, Wenhe Xu, Yang Li, Dawei Dai, Rui He, Wenhua Zhou, Jiahong Wang, Xue-Feng Yu","doi":"10.1038/s42256-026-01298-6","DOIUrl":"10.1038/s42256-026-01298-6","url":null,"abstract":"Current large language models require hundreds of billions of parameters yet struggle with domain-specific reasoning and tool coordination in materials science. Here we present MatBrain, a lightweight collaborative agent system comprising two synergistic models specialized for crystal materials research. MatBrain uses a dual-model architecture: Mat-R1 (30B parameters) as the analytical model providing expert-level domain reasoning, and Mat-T1 (14B parameters) as the executive model orchestrating tool-based actions. Entropy analysis reveals distinct output-distribution profiles for tool planning and analytical reasoning, providing a diagnostic signal consistent with the functional specialization of the two modules. Enabled by this dual-model architecture and structural efficiency, MatBrain is competitive with frontier large language models while being lightweight and locally deployable. MatBrain exhibits versatility across structure generation, property prediction and synthesis planning tasks. Applied to catalyst design, MatBrain generated 30,000 candidate structures and identified 38 promising materials within 48 h, while significantly reducing the human-active time required for materials design and computational screening. These results demonstrate the potential of lightweight collaborative intelligence for advancing materials research capabilities. Shi et al. demonstrate a dual architecture approach for materials research, integrating two lightweight large language models for collaborative reasoning and scientific tool execution. The method achieves competitive performance while remaining affordable and locally deployable.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 9","pages":"1441-1456"},"PeriodicalIF":29.8,"publicationDate":"2026-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148932532","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}

引用次数: 0

Causal evidence that language models use confidence to drive behaviour 因果证据表明，语言模型使用信心来驱动行为

IF 29.8 1区计算机科学

Nature Machine Intelligence Pub Date : 2026-09-07 DOI: 10.1038/s42256-026-01293-x

Dharshan Kumaran, Nathaniel Daw, Simon Osindero, Petar Veličković, Viorica Patraucean

{"title":"Causal evidence that language models use confidence to drive behaviour","authors":"Dharshan Kumaran, Nathaniel Daw, Simon Osindero, Petar Veličković, Viorica Patraucean","doi":"10.1038/s42256-026-01293-x","DOIUrl":"10.1038/s42256-026-01293-x","url":null,"abstract":"Metacognition—assessing the quality of one’s own cognitive performance—guides adaptive behaviour across species. Confidence signals can be extracted from language model outputs, yet a fundamental question remains: do models actually use these signals to decide whether to answer or abstain? Here we developed a four-phase paradigm. Phase 1 elicited baseline confidence without an abstention option. Phase 2 showed that large language models apply an implicit threshold to internal confidence when abstaining, with confidence effect sizes roughly an order of magnitude larger than alternative mechanisms. Phase 3 provided causal evidence via activation steering: boosting or suppressing confidence correspondingly decreased or increased abstention, with mediation analysis confirming confidence redistribution as the primary mechanism. Phase 4 instructed models to abstain at different confidence levels and showed they adjusted their behaviour accordingly, indicating they read out and act on confidence to set abstention policies. Critically, verbal confidence—an explicit self-evaluation from a separate forward pass—independently predicted abstention across all models despite being less discriminatory of correctness, and activation decoding showed both measures (that is, token log probabilities and verbal confidence) are lossy read-outs of a richer internal representation. Together, these results suggest that abstention is captured by the joint operation of a multidimensional internal confidence representation and threshold-based policies—consistent with structured metacognitive control in large language models—a capacity of growing importance as models transition to autonomous agents that must recognize their own uncertainty. Kumaran et al. show that large language models making decisions on when to answer a question or abstain from answering can be influenced by boosting or suppressing confidence signals in the model.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 9","pages":"1411-1425"},"PeriodicalIF":29.8,"publicationDate":"2026-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.comhttps://www.nature.com/articles/s42256-026-01293-x.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148899594","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}

引用次数: 0

NucleicBERT interprets RNA sequence space through self-supervised language modelling nuclebert通过自我监督语言建模来解释RNA序列空间

IF 29.8 1区计算机科学

Nature Machine Intelligence Pub Date : 2026-09-03 DOI: 10.1038/s42256-026-01295-9

Utkarsh Upadhyay, Julian Herold, Markus Götz, Alexander Schug

{"title":"NucleicBERT interprets RNA sequence space through self-supervised language modelling","authors":"Utkarsh Upadhyay, Julian Herold, Markus Götz, Alexander Schug","doi":"10.1038/s42256-026-01295-9","DOIUrl":"10.1038/s42256-026-01295-9","url":null,"abstract":"Much of the human genome’s non-protein-coding fraction acts directly through RNA, yet the structural and functional roles encoded in these sequences remain poorly understood. Applying deep learning is hindered by scarce RNA structural data and it remains unclear what biological constraints such models can recover directly from the abundant RNA sequences alone. Here, to address these challenges, we developed NucleicBERT, a self-supervised masked-language model that learns contextual representations from single sequences without evolutionary information. Explainable artificial intelligence analyses show that the model organizes RNA sequences in latent space and encodes structural properties indicating that biologically meaningful constraints are learned from sequence correlations alone. When fine-tuned for downstream structural and functional tasks, NucleicBERT requires only single sequences while matching or exceeding current RNA prediction models. This alignment-free framework addresses the scarcity of annotated 3D RNA data while providing a rapid, computational complement to experimental techniques. By bridging abundant unlabelled sequence data with scarce structural annotations, NucleicBERT advances RNA structure prediction and informs how large language models encode biological information. RNA structure and function are hard to infer because annotations are scarce, despite abundant sequence data. Upadhyay et al. trained a self-supervised model on large-scale RNA data that derives biologically meaningful patterns from sequence correlations.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 9","pages":"1426-1440"},"PeriodicalIF":29.8,"publicationDate":"2026-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.comhttps://www.nature.com/articles/s42256-026-01295-9.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148894917","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}

引用次数: 0

Quantum neural operators with implicit quadratic frame and expressivity advantages 具有隐式二次框架和表达性优势的量子神经算子

IF 29.8 1区计算机科学

Nature Machine Intelligence Pub Date : 2026-09-03 DOI: 10.1038/s42256-026-01289-7

Ruocheng Wang, Xiaoqiu Zhong, Zhuo Xia, Junchi Yan

{"title":"Quantum neural operators with implicit quadratic frame and expressivity advantages","authors":"Ruocheng Wang, Xiaoqiu Zhong, Zhuo Xia, Junchi Yan","doi":"10.1038/s42256-026-01289-7","DOIUrl":"10.1038/s42256-026-01289-7","url":null,"abstract":"While classical neural operators have transformed the solving of partial differential equations, developing efficient quantum counterparts remains challenging due to processing overheads and theoretical gaps. Specifically, existing quantum machine learning paradigms often demand prohibitive qubit scaling or deep circuits, rendering them impractical for near-term hardware. Here, to bridge this gap, we propose QuanONet, a quantum neural operator tailored for the noisy intermediate-scale quantum era. Theoretically, we extend the universal approximation theorem to the quantum domain for continuous nonlinear operators. Departing from conventional views relying on exponential Hilbert spaces, we prove that the architecture’s density matrix implicitly constructs a quadratic feature frame. We establish that, for operators, this implicit frame yields an $${mathcal{O}}({p}^{2})$$ expressivity bound, circumventing the $${mathcal{O}}(p)$$ linear capacity limits of matched classical models. To overcome the depth cost of spectrum alignment, we introduce a trainable-frequency (TF-QuanONet) strategy that adaptively spaces base frequencies to capture relatively high frequencies without parameter inflation. Extensive benchmark experiments demonstrate that TF-QuanONet notably outperforms quantum baselines and achieves competitive accuracy against classical frameworks under strictly matched-parameter conditions. Furthermore, in high-dimensional scaling regimes (p → 256), the architecture exhibits superior optimization robustness, consistently converging to the intrinsic error floor where classical baselines suffer from high variance. The physical deployment on IBM quantum processors validates its functional resilience on near-term hardware as a qualitative proof of concept. Wang et al. introduce a hardware-efficient quantum neural operator that overcomes classical linear capacity limits. Using an implicit quadratic frame, it offers accelerated expressivity for solving differential equations in the noisy intermediate-scale quantum era.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 9","pages":"1371-1380"},"PeriodicalIF":29.8,"publicationDate":"2026-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148894916","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}

引用次数: 0

Steering machine reasoning with brain signals 操纵机器用大脑信号推理

IF 29.8 1区计算机科学

Nature Machine Intelligence Pub Date : 2026-09-01 DOI: 10.1038/s42256-026-01302-z

Changde Du, Huiguang He

引用次数: 0