Summary
Machine Learning (ML) is a sub-field of artificial intelligence concerned with algorithms that improve their performance through iterative exposure to data. At its core, ML involves constructing models—parametrised functions or probability distributions—that generalise patterns from input to predict outcomes on new data. Supervised learning relies on labelled examples to learn mappings between inputs and outputs, while unsupervised learning uncovers hidden structures within unlabelled datasets. Reinforcement learning agents discover optimal decision policies by trial-and-error interaction with an environment, guided by reward signals. Deep neural networks, with their multi-layered architectures and nonlinear activation functions, have driven breakthroughs in vision, language processing and scientific simulation. Progress in ML now depends equally on algorithmic advances—in optimisation, regularisation and network design—and on access to large, high-quality datasets and specialised hardware. As ML permeates fields from healthcare diagnostics to climate modelling, ensuring interpretability, fairness and robustness remains vital to its responsible deployment.
Research from Nature Portfolio
A non-parametric neural approach to meter-coefficient estimation employs a back-propagation network to model complex relationships in domestic electricity readings. By integrating bespoke feature extraction, adaptive preprocessing and network tuning, the study achieves a 40 % reduction in mean absolute error compared with traditional least-squares analysis, enabling more accurate real-time smart-grid monitoring.
In analogue deep-neural-network hardware, algorithmic noise injection has been shown to enhance resilience without hardware modification. A Bayes-guided perturbation framework injects calibrated noise into activations, delivering one to two orders of magnitude improvements in robustness across image classification, object detection and large-scale point-cloud tasks while maintaining baseline accuracy under device variability.
Transformer-based models for land-use and land-cover classification leverage transfer learning and explainability tools to balance efficiency and transparency in satellite imagery analysis. By fine-tuning pretrained encoders and employing attribution libraries, the method identifies and mitigates biases in environmental and urban planning applications, producing accurate, interpretable LULC maps at minimal computational cost.
Topic trend for the past 5 years
The graph below shows the article count in Nature Index journals for machine learning.
* The ‘Current Index’ represents data for a 12-month rolling window, the current window is 1 May 2025 - 30 April 2026.
Technical terms
Back-propagation: A gradient-based method that computes error derivatives in neural networks, propagating them from the output layer back through intermediate weights and biases to update model parameters.
Kernel function: A similarity measure that implicitly maps inputs into a high-dimensional feature space, enabling linear algorithms to perform nonlinear classification or regression.
Particle swarm optimisation: A population-based heuristic in which candidate solutions adjust their positions and velocities by combining individual best experiences with global swarm intelligence.
Self-attention: A mechanism that computes dynamic, pairwise affinities between elements in a sequence, allowing models to aggregate contextual information across long ranges without recurrence.
Transformer: A neural architecture comprising stacked self-attention and feed-forward layers, designed for flexible sequence modelling and parallelisable computation.
Notable articles in machine learning
- Mastering the game of Go with deep neural networks and tree search. Nature (2016).
- Mastering the game of Go without human knowledge. Nature (2017).
- Mastering Atari, Go, chess and shogi by planning with a learned model. Nature (2020).
- Deep learning with coherent nanophotonic circuits. Nature Photonics (2017).
- Earthquake transformer—an attentive deep-learning model for simultaneous earthquake detection and phase picking. Nature Communications (2020).
- Clinical-grade computational pathology using weakly supervised deep learning on whole slide images. Nature Medicine (2019).
About these summaries
This Nature Research Intelligence Topic summary is created with the cited references and a large language model. We take care to ground generated text with facts, and have systems in place to gain human feedback on the overall quality of the process in line with our AI principles. We strive to create accurate and useful summaries for people unfamiliar with the research topic and that supports this goal. These pages are a beta release and will be updated as we learn how best to help people gain value from a research topic summary.
Research
Position of Machine Learning in Nature Index by Count
Leading institutions
| Institution | Count | Share |
|---|---|---|
| Tsinghua University | 54 | 20.05 |
| Chinese Academy of Sciences (CAS) | 60 | 16.45 |
| Shanghai Jiao Tong University (SJTU) | 29 | 13.68 |
| Zhejiang University (ZJU) | 25 | 13.11 |
| Xi'an Jiaotong University (XJTU) | 19 | 12.47 |
| Beijing Institute of Technology (BIT) | 21 | 11.2 |
| Peking University (PKU) | 36 | 10.55 |
| The University of Hong Kong (HKU) | 20 | 9.55 |
| Tongji University | 17 | 8.62 |
| University of Pennsylvania (Penn) | 20 | 8.48 |
Leading countries/territories
| Countries/territories | Count | Share |
|---|---|---|
| China | 415 | 369.86 |
| United States of America (USA) | 277 | 198.54 |
| United Kingdom (UK) | 93 | 45.38 |
| Germany | 71 | 38.11 |
| South Korea | 41 | 35.87 |
| Japan | 31 | 17.32 |
| Switzerland | 29 | 16.73 |
| Australia | 31 | 16.11 |
| France | 24 | 13.61 |
| Canada | 28 | 13.07 |
Collaboration
Top 5 leading collaborators in Machine Learning
Collaborating institutions
Note: Hover over the bars to view details about each institution's Share.
Turn complex research questions into confident strategic decisions
When you're under pressure to set direction, justify investment, or understand your competitive position, you need more than raw data — you need trusted insights you can act on.
Benchmark your performance against global peers using robust, methodologically sound analysis.
Combine quantitative metrics with qualitative expert insight to uncover strengths, gaps and emerging opportunities.
Gain tailored, decision-ready recommendations aligned to your strategic priorities.
Talk to us to learn more about our data dashboards and bespoke strategy reports.
Grow research skills, confidence and careers with training built for every stage of the research lifecycle.
Developed with Nature Portfolio journal Editors and internationally renowned experts. Discover three ways to learn:
Self-paced, online courses in convenient bite-sized units, covering key skills across scientific writing, publishing, grant writing, data analysis, and more.
Expert trainer-led workshops with hands-on exercises and real-time feedback across core research skills, delivered via interactive group sessions.
Editor-led workshops combining core principles in writing and publishing, personalised 1:1 feedback from Nature Portfolio Editors and hands-on exercises.
Explore course catalogues and workshop agendas, enquire about the options or request institutional pricing.