Nature Machine Intelligence最新文献

{"title":"A dexterous soft hand exoskeleton restores intentional grasping in individuals with severe hand impairment","authors":"John Nassour, Nicolas Berberich, Daniel Utpadel-Fischler, Tobias Wächter, Gordon Cheng","doi":"10.1038/s42256-026-01263-3","DOIUrl":"10.1038/s42256-026-01263-3","url":null,"abstract":"Soft hand exoskeletons have emerged as promising assistive devices for individuals with impaired hand function. However, most existing systems provide limited dexterity and primarily target users with moderate hand ability, leaving individuals with severe hand paralysis without effective solutions for reliable grasping of diverse objects. Here we report the translational development of a lightweight, textile-based soft robotic exoskeleton glove with wrist dorsiflexion and an active opposable and abductable thumb, designed to restore hand function in a patient with severe right-hand impairment due to amyotrophic lateral sclerosis. We followed a co-creation approach, enhancing dexterity by increasing hand articulations based on patient needs. Furthermore, to enhance the patient’s sense of control, a non-invasive surface electromyography-based grasp predictor (97% sensitivity) was combined with motion data and machine learning-based error correction to compensate for weak, noisy muscle signals, compared with healthy controls (n = 15). The exoskeleton enabled the patient to grasp objects, achieve a Box-and-Blocks Test score of 5 and perform meaningful tasks, including feeding himself. We further validated the exoskeleton in patients with stroke (n = 6). While exoskeleton assistance on average reduced Action Research Arm Test scores of moderately impaired patients by 9, severely impaired patients scored 17 points higher when using the exoskeleton. These results indicate that the dexterous soft hand exoskeleton is particularly effective for individuals with severe to near-complete hand paralysis, while its utility for patients with moderate residual function is limited and task dependent. Nassour, Berberich and colleagues present a soft robotic hand exoskeleton that restores grasping ability in individuals with severe hand paralysis, enabling meaningful tasks such as feeding. A lightweight textile glove with wrist dorsiflexion and an active opposable thumb increases hand articulations to enable more dexterous grasping.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 7","pages":"1100-1114"},"PeriodicalIF":29.8,"publicationDate":"2026-06-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.comhttps://www.nature.com/articles/s42256-026-01263-3.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148305223","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":"Autonomous navigation of intelligent microrobotic swarms in unknown environments","authors":"Xuanyu An, Shengming Luo, Haoyu Zhang, Qijun Yang, Yuanbiao Ma, Bin Wang, Jiansheng Du, Qianqian Wang","doi":"10.1038/s42256-026-01252-6","DOIUrl":"10.1038/s42256-026-01252-6","url":null,"abstract":"Microrobotic swarms, through their collective and reconfigurable behaviours, can flexibly adapt to various targeted delivery and navigation tasks. However, achieving autonomous navigation and obstacle avoidance in unknown environments remains a challenge. Here we propose a reinforcement-learning-based control strategy for swarm navigation under partial observation. To bridge the gap between simulation and reality, we introduce a generalized sim-to-real transfer method, enabling the simulation-trained swarm to effectively explore in unknown environments. Our model combines temporally extended attention with multi-level domain randomization of the environment, perception, actuation and dynamics, allowing the policy to use current sensory inputs and historical context to select magnetic actuation commands. Our model efficiently performs autonomous navigation and obstacle avoidance through benchmarking against human operators in simulated environments. The proposed strategy enables swarm navigation and dynamic obstacle avoidance, cargo transportation, moving target tracking and recovery from temporary vision loss, and hovering by its partial sensing and decision-making. Analysis of action sequences and attention scores reveals that the swarm makes task-priority-based decisions, optimizing trajectories towards the target in unknown environments. An, Luo, Zhang and colleagues present Turbo, a transformer-based reinforcement learning framework that enables simulation-to-real transfer for autonomous navigation and obstacle avoidance in physical microrobotic swarms operating in unknown environments.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 6","pages":"955-968"},"PeriodicalIF":29.8,"publicationDate":"2026-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148290818","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":"Algorithm–hardware co-design of neuromorphic networks with dual memory pathways","authors":"Pengfei Sun, Zhe Su, Jascha Achterberg, Giacomo Indiveri, Dan F. M. Goodman, Danyal Akarca","doi":"10.1038/s42256-026-01255-3","DOIUrl":"10.1038/s42256-026-01255-3","url":null,"abstract":"Spiking neural networks excel at event-driven sensing. Yet, maintaining task-relevant context over long timescales both algorithmically and in hardware, while respecting both tight energy and memory budgets, remains a core challenge in the field. Here we address this challenge through an algorithm–hardware co-design effort. At the algorithm level, inspired by the cortical fast–slow organization in the brain, we introduce a neural network with an explicit slow memory pathway that, combined with fast spiking activity, enables a dual memory pathway architecture in which each layer maintains a compact low-dimensional state that summarizes recent activity and modulates spiking dynamics. This explicit memory stabilizes learning while preserving event-driven sparsity, achieving competitive accuracy on long-sequence benchmarks with 40–60% fewer parameters than equivalent state-of-the-art spiking neural networks. At the hardware level, we introduce a near-memory-compute architecture that fully leverages the advantages of the dual memory pathway architecture by retaining its compact shared state while optimizing data flow, across heterogeneous sparse-spike and dense-memory pathways. We show experimental results that demonstrate more than a fourfold increase in throughput and over a fivefold improvement in energy efficiency compared with state-of-the-art implementations. Together, these contributions demonstrate that biological principles can guide functional abstractions that are both algorithmically effective and hardware-efficient, establishing a scalable co-design framework for real-time neuromorphic computation and learning. Pengfei Sun et al. develop a spiking neural network with a dual memory pathway, co-designed with a custom neuromorphic chip. The approach delivers over 4× throughput and 5x energy efficiency gains while using 40–60% fewer parameters than state-of-the-art implementations.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 6","pages":"901-912"},"PeriodicalIF":29.8,"publicationDate":"2026-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.comhttps://www.nature.com/articles/s42256-026-01255-3.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148290817","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":"Towards AI-augmented decision making in psychiatry","authors":"Jiyeong Kim, Carolyn I. Rodriguez","doi":"10.1038/s42256-026-01256-2","DOIUrl":"10.1038/s42256-026-01256-2","url":null,"abstract":"Psychiatric disorders are heterogeneous, and care depends on interpreting unstructured longitudinal narratives, creating variability that hinders standardization. A study now shows that a psychiatry-specific large language model (LLM) may help clinicians to deliver more consistent, high-quality care.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 6","pages":"864-865"},"PeriodicalIF":29.8,"publicationDate":"2026-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148300152","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":"From virtual experiments to biomedical insight with synthetic data","authors":"Margarida Victoriano, Milena Pavlović, Geir Kjetil Sandve, Hélder P. Oliveira, Artur Rocha, Victor Greiff","doi":"10.1038/s42256-026-01244-6","DOIUrl":"10.1038/s42256-026-01244-6","url":null,"abstract":"Synthetic datasets are essential for the development and benchmarking of machine learning methods in biomedicine, as they help overcome the pervasive data scarcity in biomedical research. In fields such as immunomics, genomics and proteomics, they enable the development of prediction algorithms, including methods for immune receptor–antigen binding prediction. When generated with transparent and fully specified parameters, synthetic datasets serve as rule-based systems for reproducible and interpretable model testing, an essential step towards digital twins that emulate biological systems for diagnosis and therapy design. A key obstacle, however, is the ‘simulation to reality’ (sim2real) gap, which describes the uncertainty about whether performance on synthetic data is predictive of performance on experimental data. Divergent statistical and biological properties may erode generalizability and clinical relevance. The lack of standardized sim2real benchmarks impedes validation and widespread adoption. We argue that multilayered validation frameworks, incorporating techniques such as domain adaptation and hybrid validation, and grounded in biological realism, are essential to ensuring that synthetic datasets faithfully capture biological complexity. Closing the sim2real gap will unlock the full translational potential of synthetic data, accelerating diagnostic and therapeutic discovery, guiding clinical decision-making, and advancing the development of predictive digital twins. Synthetic datasets are becoming crucial for the development of biomedical machine learning models. Victoriano et al. discuss the persistent simulation-to-reality gap that limits how well synthetic performance predicts real-world performance.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 6","pages":"866-879"},"PeriodicalIF":29.8,"publicationDate":"2026-06-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148300148","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":"Bridging three-dimensional molecular structures and artificial intelligence with a conformation description language","authors":"Jiacheng Xiong, Yuqi Shi, Min Wu, Panpan Shao, Zhaokun Wang, Wei Zhang, Runze Zhang, Zhiyi Chen, Chuanlong Zeng, Xun Jiang, Duanhua Cao, Zhaoping Xiong, Zunyun Fu, Sulin Zhang, Mingyue Zheng","doi":"10.1038/s42256-026-01250-8","DOIUrl":"10.1038/s42256-026-01250-8","url":null,"abstract":"Large language models have revolutionized many fields by learning from sequential data, yet their application to three-dimensional molecular modelling has been hindered by the lack of effective token-based representations of molecular conformations. Here we present ConfSeq, a conformation description language that addresses this gap by encoding three-dimensional molecular structures into discrete token sequences. ConfSeq integrates molecular SMILES with internal coordinates, including dihedral angles, bond angles and a pseudo-chirality descriptor, thereby ensuring SE(3) invariance and preserving the conciseness and human readability inherent to SMILES. By reformulating core three-dimensional molecular modelling tasks (including conformation prediction, de novo generation and representation learning) as sequence modelling problems, ConfSeq enables standard transformer architectures to achieve state-of-the-art performance across diverse benchmarks. Furthermore, ConfSeq enabled the discovery of multiple novel stimulator of interferon gene (STING) inhibitors and ALDH1B1 inhibitors with half-maximal inhibitory concentrations of 0.338–3.51 μM. Collectively, these findings establish ConfSeq as a robust framework for extending large language models in three-dimensional molecular modelling. Xiong et al. introduce ConfSeq, a molecular conformation description language that enables language models to perform three-dimensional molecular modelling tasks, including conformer prediction, three-dimensional molecular generation and representation, with strong performance.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 6","pages":"969-983"},"PeriodicalIF":29.8,"publicationDate":"2026-06-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148288843","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":"Reusability report: Assessment of reproducibility and applicability to external datasets for RXNGraphormer","authors":"Chonghuan Zhang, Majian Zeng, Qianghua Lin, Kuangbiao Liao","doi":"10.1038/s42256-026-01257-1","DOIUrl":"10.1038/s42256-026-01257-1","url":null,"abstract":"Deep learning has substantially advanced reaction-yield prediction and synthesis-planning methodologies, yet achieving a unified architecture capable of transferring across these tasks remains a central challenge in chemical machine learning. Xu et al. recently developed RXNGraphormer, which combines a pretrained graph–transformer encoder with a delta-molecular reaction representation designed to support cross-task generalization. Here, in this reusability report, we independently assess the reproducibility and practical applicability of RXNGraphormer using the released implementation, pretrained checkpoint and benchmark datasets. All major regression and sequence-generation results reported in the original study were consistently reproduced, including the relative difficulty patterns in out-of-sample evaluations, demonstrating the stability and transparency of the published workflow. To evaluate reusability, we examined the model’s transfer to multiple high-throughput datasets generated under standardized experimental conditions. In these settings, the pretrained encoder adapted efficiently and delivered strong predictive performance with minimal fine-tuning. On an external sequence-prediction benchmark beyond the original USPTO setting, RXNGraphormer retained strong forward prediction capability, whereas retrosynthetic prediction was more sensitive to distributional shift. Overall, our results show that RXNGraphormer is a reproducible and practically reusable framework for unified reaction learning, while also highlighting the continued importance of harmonized reaction representations, curated data and domain-specific refinement. Based on independent reproduction and external benchmarks, this study shows that RXNGraphormer is reproducible and transferable across reaction prediction tasks, while revealing limitations under heterogeneous and distribution-shifted settings.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 7","pages":"1066-1074"},"PeriodicalIF":29.8,"publicationDate":"2026-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148288874","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":"Navigating molecular OOD-ness","authors":"Anna Tomberg, Nessa Carson","doi":"10.1038/s42256-026-01251-7","DOIUrl":"10.1038/s42256-026-01251-7","url":null,"abstract":"Machine learning methods for drug discovery often face difficulties in identifying novel bioactive molecules that do not belong to the training data distribution. A new metric can now quantify chemical distribution shift and evaluate the generalization capability of molecular machine learning models.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 6","pages":"862-863"},"PeriodicalIF":29.8,"publicationDate":"2026-06-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148300153","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":"Explicit dynamic cross-strand interactions for DNA sequence language modelling","authors":"Cheng Yang, Yuansheng Liu, Lei Ling, Yang Liu, Fengxin Li, Changjian Chen, Long Wang, Feng Yu, Liang Qiao, Xiangxiang Zeng, Kenli Li, Alexander Schönhuth, Xiao Luo","doi":"10.1038/s42256-026-01249-1","DOIUrl":"10.1038/s42256-026-01249-1","url":null,"abstract":"Large language models are transforming the functional interpretation of genomic sequences. Most DNA sequence language models either treat DNA as unidirectional text or enforce reverse-complement symmetry through data augmentation or equivariance, yet both approximate double-strand interactions implicitly and statically, limiting the capture of context-dependent cross-strand information exchange during representation learning. In biological systems, double-strand coordination is shaped by physical coupling, functional synergy and information transfer, and is fundamental to genomic function. Here we present CrossDNA, an explicit and dynamic language model for DNA cross-strand modelling. CrossDNA uses dual branches with alternating exposure to forward and reverse-complement sequence segments as a duplex-inspired training strategy, and establishes interstrand communication through a lightweight cross-strand communication module. To support long-context modelling, CrossDNA couples a recurrent long-context backbone with sliding-window attention. Across diverse genomics tasks, CrossDNA delivers consistent performance gains and improves robustness to sequence orientation, particularly in enhancer prediction. CrossDNA uses only million-scale parameters yet achieves comparable or superior performance to DNA foundation models containing hundreds of millions of parameters, demonstrating markedly greater parameter efficiency. These results advance DNA representation from the implicit, static approximation of double-strand relationships to explicit, dynamic cross-strand modelling, enabling regulatory logic interpretation, discovery of novel regulatory elements and prioritization of disease-associated non-coding variants. Yang et al. developed CrossDNA, a parameter-efficient language model for cross-strand DNA modelling inspired by double-strand dynamics. It performs strongly in benchmarks and supports regulatory region interpretation and non-coding variant prioritization.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 6","pages":"880-900"},"PeriodicalIF":29.8,"publicationDate":"2026-06-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148290477","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":"Conditional Monge Gap enables generalizable single-cell perturbation modelling","authors":"Alice Driessen, Dhruva Abhijit Rajwade, Benedek Harsanyi, Marianna Rapsomaniki, Jannis Born","doi":"10.1038/s42256-026-01242-8","DOIUrl":"10.1038/s42256-026-01242-8","url":null,"abstract":"Learning the response of single cells to various treatments offers great potential to enable targeted therapies. In this context, neural optimal transport has emerged as a principled methodological framework because it inherently accommodates the challenges of unpaired data induced by cell destruction during data acquisition. However, most existing optimal transport approaches are incapable of conditioning on different treatment contexts (for example, time, drug treatment, drug dose or cell type), and we still lack methods that unanimously show promising generalizability to unseen treatments. Here we propose the Conditional Monge Gap (CMonge), which learns optimal transport maps conditionally on arbitrary covariates. We demonstrate its value in predicting single-cell perturbation responses conditional to one or more drugs, drug dose or combinations thereof. We found that our conditional models achieve results comparable with and sometimes even superior to the condition-specific state-of-the-art single-cell RNA sequencing as well as multiplexed protein imaging data. Notably, by scaling to hundreds of conditions and training on hundreds of millions of drugs, we enable cross-task learning and unlock generalizability to unseen drugs. Our method widely outperforms other conditional models in capturing heterogeneity in cell populations. In short, CMonge is mathematically grounded, highly parameter-efficient relative to single-cell foundation models and yields accurate predictions for unseen drugs using only the compound structure. Thus, it opens a practical route for accelerating drug discovery and repurposing. Driessen et al. present a conditional optimal transport method that can model the distribution shift between perturbed and unperturbed cell transcriptomes and that can generalize to unseen contexts.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"8 6","pages":"984-996"},"PeriodicalIF":29.8,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.comhttps://www.nature.com/articles/s42256-026-01242-8.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148300156","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}