Summary
Artificial Intelligence (AI) encompasses the design of computational systems capable of performing tasks that normally require human intelligence. Grounded in interdisciplinary insights from computer science, neuroscience and linguistics, AI divides broadly into symbolic approaches—where knowledge is encoded in formal rules—and connectionist or learning-based approaches, which employ artificial neural networks to discover patterns from data. Over recent decades, advances in machine learning and, in particular, deep learning have enabled breakthroughs in perception, natural language understanding and complex decision-making. AI systems now analyse medical images for early detection of disease, translate speech in real time, drive autonomous vehicles and personalise recommendations on a global scale. Key challenges include ensuring robust performance under distribution shifts, interpretability of model decisions, data privacy and the integration of reasoning with learning. As compute power and data availability continue to expand, AI promises to underpin major advances in science, healthcare, manufacturing and environmental stewardship.
Research from Nature Portfolio
A new visual-speech recognition architecture has been shown to outperform much larger prior models on lip-reading benchmarks across multiple languages. By introducing auxiliary prediction tasks alongside the primary lip-reading objective and optimising hyperparameters and data augmentations, the system closes much of the gap between video-only and audio-based recognition.
Researchers have developed flexible, self-powered triboelectric sensors inspired by the eardrum to capture lip movements as paired phase-shifted vibration signals. These signals are converted into cross-recurrence plots and fed to a dilated recurrent neural network trained with a prototype learning objective. The combination replaces conventional spectrogram inputs and achieves competitive word-classification accuracy with lower computational cost.
A two-stage lip-reading framework tailored for speech-impaired intensive-care patients first predicts intermediate acoustic features from facial frames, then decodes these into text. Trained and evaluated on a bespoke ICU corpus, the model achieves word-error rates below 7 percent, offering a practical communication aid for tracheotomised individuals.
Topic trend for the past 5 years
The graph below shows the article count in Nature Index journals for artificial intelligence.
* The ‘Current Index’ represents data for a 12-month rolling window, the current window is 1 May 2025 - 30 April 2026.
Technical terms
Visual-speech recognition: The task of converting silent lip movements or facial video into textual transcriptions without relying on audio input.
Triboelectric sensor: A flexible, self-powered device that generates electrical signals in response to mechanical deformation, here used to capture fine lip motions.
Dilated recurrent neural network: A sequence model in which recurrent connections are “dilated” to capture long-range temporal dependencies with fewer layers.
Prototype learning: A metric-based training objective in which samples are compared to learned class prototypes in feature space, improving generalisation with limited data.
Auxiliary loss: An additional training objective imposed on intermediate layers of a network to encourage the learning of useful features alongside the primary task.
Multimodal fusion: The integration of information from multiple sensor modalities—such as audio, video and motion—to improve robustness and accuracy of AI systems.
Notable articles in artificial intelligence
- A deep-learning system bridging molecule structure and biomedical text with comprehension comparable to human professionals. Nature Communications (2022).
- Pushing the limits of remote RF sensing by reading lips under the face mask. Nature Communications (2022).
- Transforming machine translation: a deep learning system reaches news translation quality comparable to human professionals. Nature Communications (2020).
- Decoding lip language using triboelectric sensors with deep learning. Nature Communications (2022).
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 Artificial Intelligence in Nature Index by Count
Leading institutions
| Institution | Count | Share |
|---|---|---|
| Shanghai Jiao Tong University (SJTU) | 12 | 7.17 |
| Tsinghua University | 18 | 6.76 |
| Zhejiang University (ZJU) | 12 | 5.95 |
| Hunan University (HNU) | 7 | 5.29 |
| Shandong University (SDU) | 7 | 5.13 |
| Chinese Academy of Sciences (CAS) | 11 | 4.38 |
| Beijing Institute of Technology (BIT) | 8 | 4.23 |
| George Mason University (GMU) | 5 | 4.16 |
| Sun Yat-sen University (SYSU) | 5 | 3.65 |
| Purdue University | 4 | 3.49 |
Leading countries/territories
| Countries/territories | Count | Share |
|---|---|---|
| China | 137 | 124.28 |
| United States of America (USA) | 79 | 61.98 |
| Germany | 25 | 17.66 |
| South Korea | 15 | 10.42 |
| Japan | 10 | 8.29 |
| United Kingdom (UK) | 15 | 6.49 |
| Switzerland | 14 | 6.21 |
| Canada | 8 | 5.4 |
| France | 9 | 5.35 |
| Netherlands | 8 | 5.06 |
Collaboration
Top 5 leading collaborators in Artificial Intelligence
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.