NATURE MACHINE INTELLIGENCE 将在机器学习、机器人和人工智能的广泛主题中发表高质量的原创研究和评论。该杂志还将探讨和讨论这些领域开始对其他科学学科以及社会和工业的许多方面产生的重大影响。在科学发现、医疗保健、医学诊断以及安全和可持续的城市、交通和农业等领域，机器智能可以增强人类能力和知识的机会不计其数。与此同时，出现了许多关于道德、社会和法律问题的重要问题，特别是考虑到快速发展的自然机器智能将提供一个平台来讨论这些广泛的影响——鼓励跨学科对话——通过评论、新闻专题、新闻和观点文章以及通信。

Nature Machine Intelligence will publish high-quality original research and reviews on a wide range of topics in machine learning, robotics, and artificial intelligence. The journal will also explore and discuss the significant impacts these fields are beginning to have on other scientific disciplines as well as many aspects of society and industry. The opportunities for machine intelligence to enhance human capabilities and knowledge are countless, particularly in areas such as scientific discovery, healthcare, medical diagnostics, and safe and sustainable cities, transportation, and agriculture. At the same time, many important ethical, social, and legal issues have emerged, especially considering the rapid development of natural machine intelligence. The journal will provide a platform to discuss these broad impacts, encouraging interdisciplinary dialogue through reviews, news features, news and opinion articles, and communications.

《Nature Machine Intelligence》将发表关于机器学习、机器人和人工智能等广泛主题的高质量原创研究和评论。该期刊还将探讨这些领域对其他科学学科、社会和工业的重大影响。在科学发现、医疗保健、医学诊断，以及安全和可持续的城市、交通和农业等领域，机器智能为增强人类能力和知识提供了无数机会。同时，随着自然机器智能的快速发展，许多关于伦理、社会和法律的问题也随之出现。该期刊将为讨论这些广泛影响提供一个平台，鼓励跨学科的对话，包括评论、新闻专题、新闻和观点文章以及通信。

Nature Machine Intelligence

In recent years, neural-network quantum states have emerged as powerful tools for the study of quantum many-body systems. Electronic structure calculations are one such canonical many-body problem that have attracted sustained research efforts spanning multiple decades, whilst only recently being attempted with neural-network quantum states. However, the complex non-local interactions and high sample complexity are substantial challenges that call for bespoke solutions. Here, we parameterize the electronic wavefunction with an autoregressive neural network that permits highly efficient and scalable sampling, whilst also embedding physical priors reflecting the structure of molecular systems without sacrificing expressibility. This allows us to perform electronic structure calculations on molecules with up to 30 spin orbitals—at least an order of magnitude more Slater determinants than previous applications of conventional neural-network quantum states—and we find that our ansatz can outperform the de facto gold-standard coupled-cluster methods even in the presence of strong quantum correlations. With a highly expressive neural network for which sampling is no longer a computational bottleneck, we conclude that the barriers to further scaling are not associated with the wavefunction ansatz itself, but rather are inherent to any variational Monte Carlo approach. To perform electronic structure calculations in quantum chemistry systems, methods are needed that are both accurate and scalable as the size of the molecule of interest increases. Barrett and colleagues employ an autoregressive neural-network ansatz that allows them to study larger molecules than previously attempted with neural-network quantum state approaches.

Thomas D. Barrett

Aleksei Malyshev

A. I. Lvovsky

Nat Mach Intell

Autoregressive neural-network wavefunctions for ab initio quantum chemistry

自回归神经网络波函数用于从头算量子化学

A central goal of artificial intelligence in high-stakes decision-making applications is to design a single algorithm that simultaneously expresses generalizability by learning coherent representations of their world and interpretable explanations of its dynamics. Here, we combine brain-inspired neural computation principles and scalable deep learning architectures to design compact neural controllers for task-specific compartments of a full-stack autonomous vehicle control system. We discover that a single algorithm with 19 control neurons, connecting 32 encapsulated input features to outputs by 253 synapses, learns to map high-dimensional inputs into steering commands. This system shows superior generalizability, interpretability and robustness compared with orders-of-magnitude larger black-box learning systems. The obtained neural agents enable high-fidelity autonomy for task-specific parts of a complex autonomous system. Inspired by the brain of the roundworm Caenorhabditis elegans , the authors design a highly compact neural network controller directly from raw input pixels. Compared with larger networks, this compact controller demonstrates improved generalization, robustness and interpretability on a lane-keeping task.

Mathias Lechner

Ramin Hasani

Alexander Amini

Thomas A. Henzinger

Daniela Rus

Radu Grosu

Neural circuit policies enabling auditable autonomy

神经回路策略实现可审计的自主性

Brent Mittelstadt

Principles alone cannot guarantee ethical AI

仅靠原则无法保证人工智能的道德性

The binding complexes formed by proteins and small molecule ligands are ubiquitous and critical to life. Despite recent advancements in protein structure prediction, existing algorithms are so far unable to systematically predict the binding ligand structures along with their regulatory effects on protein folding. To address this discrepancy, we present NeuralPLexer, a computational approach that can directly predict protein-ligand complex structures solely using protein sequence and ligand molecular graph inputs. NeuralPLexer adopts a deep generative model to sample the 3D structures of the binding complex and their conformational changes at an atomistic resolution. The model is based on a diffusion process that incorporates essential biophysical constraints and a multi-scale geometric deep learning system to iteratively sample residue-level contact maps and all heavy-atom coordinates in a hierarchical manner. NeuralPLexer achieves state-of-the-art performance compared to all existing methods on benchmarks for both protein-ligand blind docking and flexible binding site structure recovery. Moreover, owing to its specificity in sampling both ligand-free-state and ligand-bound-state ensembles, NeuralPLexer consistently outperforms AlphaFold2 in terms of global protein structure accuracy on both representative structure pairs with large conformational changes (average TM-score=0.93) and recently determined ligand-binding proteins (average TM-score=0.89). Case studies reveal that the predicted conformational variations are consistent with structure determination experiments for important targets, including human KRAS$^\textrm{G12C}$, ketol-acid reductoisomerase, and purine GPCRs. Our study suggests that a data-driven approach can capture the structural cooperativity between proteins and small molecules, showing promise in accelerating the design of enzymes, drug molecules, and beyond.

Zhuoran Qiao

Weili Nie

Arash Vahdat

Thomas F. Miller

Animashree Anandkumar

State-specific protein–ligand complex structure prediction with a multiscale deep generative model

多尺度深度生成模型用于特定状态蛋白质-配体复合物结构预测

It is widely known that neural networks (NNs) are universal approximators of continuous functions. However, a less known but powerful result is that a NN with a single hidden layer can accurately approximate any nonlinear continuous operator. This universal approximation theorem of operators is suggestive of the structure and potential of deep neural networks (DNNs) in learning continuous operators or complex systems from streams of scattered data. Here, we thus extend this theorem to DNNs. We design a new network with small generalization error, the deep operator network (DeepONet), which consists of a DNN for encoding the discrete input function space (branch net) and another DNN for encoding the domain of the output functions (trunk net). We demonstrate that DeepONet can learn various explicit operators, such as integrals and fractional Laplacians, as well as implicit operators that represent deterministic and stochastic differential equations. We study different formulations of the input function space and its effect on the generalization error for 16 different diverse applications. Neural networks are known as universal approximators of continuous functions, but they can also approximate any mathematical operator (mapping a function to another function), which is an important capability for complex systems such as robotics control. A new deep neural network called DeepONet can lean various mathematical operators with small generalization error.

Lu Lu

Pengzhan Jin

Guofei Pang

Zhongqiang Zhang

George Em Karniadakis

Learning nonlinear operators via DeepONet based on the universal approximation theorem of operators

Abbreviation	Nat Mach Intell
Journal Impact	18.61
Quartiles(Global)	COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS(Q1)

ISSN	2522-5839
h-index	67
Top Journals	Yes

Publication Information	Publisher: Springer International Publishing，Publishing cycle: 12 issues per year，Journal Type: journal，Open Access Journals: No
Basic data	Year of publication: 2019，Proportion of original research papers: 95.61%，Self Citation Rate:2.70%， Gold OA Rate: 20.52%

Nature Machine Intelligence

Journal Citation Format

A journal article with 1 author

A journal article with 2 authors

A journal article with 3 authors

A journal article with 5 or more authors

Books Citation Format

Thesis Citation Format

Web sites Citation Format

Patent Citation Format

Staying up late manually editing references? ivySCI automatically matches journals and helps you generate references with a single click.

Share Submission Experience

Autoregressive neural-network wavefunctions for ab initio quantum chemistry

Neural circuit policies enabling auditable autonomy

Principles alone cannot guarantee ethical AI

State-specific protein–ligand complex structure prediction with a multiscale deep generative model

Learning nonlinear operators via DeepONet based on the universal approximation theorem of operators

Kernel approximation using analogue in-memory computing

Envisioning better benchmarks for machine learning PDE solvers

Discussions of machine versus living intelligence need more clarity

Stable Cox regression for survival analysis under distribution shifts

Reply to: Deeper evaluation of a single-cell foundation model