Atlas report
Who works on Adversarial Robustness in Machine Learning, and where
Relative to its size, Singapore leans hardest on Adversarial Robustness in Machine Learning: 3.2% of its subject assignments, 1.7 times the world share.
The short answer
24 271 indexed researchers with a known country work on Adversarial Robustness in Machine Learning; the subject takes 1.9% of all subject assignments in the index.
Relative to its size, Singapore leans hardest on Adversarial Robustness in Machine Learning: 3.2% of its subject assignments, 1.7 times the world share.
Since 2024, Adversarial Robustness in Machine Learning drew 10 427 newcomers against 4 110 in the previous window, a ratio of 2.5 where the average subject sits at 1.6.
What stands out
Among doctoral starts, Adversarial Robustness in Machine Learning takes 1.8 times the share it took over the previous three years, +22 878 people.
In 2025, the last full year, Adversarial Robustness in Machine Learning gained 1.2 times as many new researchers as in an average year of the three before; the world figure is 0.8.
By country
China and the United States hold 56% of the people working on Adversarial Robustness in Machine Learning, more than the 53% the same two countries hold across all subjects.
Country reports: China, United States, Germany, United Kingdom, India, South Korea, France.
| Country | Researchers | Share of the subject | Share of the country's researchers |
|---|---|---|---|
| China | 6 836 | 28% | 1.7% |
| United States | 6 643 | 27% | 2.3% |
| Germany | 1 204 | 5.0% | 2.0% |
| United Kingdom | 1 127 | 4.6% | 2.2% |
| India | 719 | 3.0% | 1.6% |
| Canada | 698 | 2.9% | 2.2% |
| South Korea | 563 | 2.3% | 1.5% |
| Australia | 554 | 2.3% | 2.7% |
| Singapore | 521 | 2.1% | 3.2% |
| France | 513 | 2.1% | 1.5% |
| Hong Kong | 451 | 1.9% | 2.2% |
| Italy | 374 | 1.5% | 1.5% |
Newcomers by year
594 researchers on Adversarial Robustness in Machine Learning declare a doctorate in progress on their ORCID record, started and without an end date yet.
On 796 dated profiles, the median time from first paper to defence on Adversarial Robustness in Machine Learning is 5.0 years.
| First publication | New on Adversarial Robustness in Machine Learning | Share of the world's newcomers |
|---|---|---|
| 2019 | 946 | 4.9% |
| 2020 | 1 045 | 5.1% |
| 2021 | 1 209 | 5.0% |
| 2022 | 1 530 | 4.6% |
| 2023 | 1 371 | 4.6% |
| 2024 | 1 314 | 4.5% |
| 2025 | 1 683 | 7.0% |
Compared with subjects of the same size
Set against Reinforcement Learning in Robotics and Advanced Vision and Imaging, two subjects of similar headcount, the difference is the pace of entry.
| Adversarial Robustness in Machine Learning | Reinforcement Learning in Robotics | Advanced Vision and Imaging | |
|---|---|---|---|
| Researchers | 24 271 | 24 353 | 24 008 |
| Largest country | China | China | China |
| Newcomers, recent against previous window | 2.5 | 2.3 | 0.9 |
Method
Subjects follow the OpenAlex taxonomy, on the profiles that carry one; a person can carry several, so shares are shares of subject assignments. A country is where the institution sits. Newcomers are dated by first publication, and a doctorate in progress is a junior profile with a first publication and no dated defence. The index is built from arXiv, OpenAlex, ORCID and DBLP and grows every night. See the atlas, how to find and hire AI researchers and how long an AI doctorate takes.
Common questions
- How many researchers work on Adversarial Robustness in Machine Learning?
- 24 271 researchers in the Founding Hires index carry Adversarial Robustness in Machine Learning as a subject, across the largest countries, as of September 2026. Worldwide the subject holds 1.9% of all subject assignments.
- Which countries lead on Adversarial Robustness in Machine Learning?
- China (6 836), United States (6 643), Germany (1 204).
- Is Adversarial Robustness in Machine Learning growing?
- Yes, faster than the world. Since 2024, 10 427 people in Adversarial Robustness in Machine Learning published for the first time, against 4 110 over the three years before, a ratio of 2.5 where the world sits at 1.6.
- How many doctorates on Adversarial Robustness in Machine Learning are in progress?
- 594 researchers on Adversarial Robustness in Machine Learning declare a doctorate in progress on their ORCID record, started and without an end date yet.
The names behind the numbers
Every count on this page is a list in the index.
Search it by subject and country to find who leads the work today, and set an alert on the ones about to come free, for $179 a month or 149 € in the euro area. Or hand the search over and pay only when someone joins.