Scientists are ‘tricked’ by artificial data generated by generative AI, but automated data analysis is getting closer
AI technology that was previously used to simulate human art is now able to synthesize artificial scientific information, which advances efforts towards fully automated data analysis.
Researchers from the University of Illinois Urbana-Champaign developed an AI to generate artificial data using microscopy experiments that are commonly used to characterize material structures at atomic level. The AI, which draws on the same technology as art generators allows researchers to include background noise and experiment imperfections in the generated data. This allows material features to detected faster and more effectively than before.
The journal npj Computational Materials published the study \”Leveraging generative antagonistic networks to create realistic scan transmission electron microscopy imaging.\”
Source:
https://techxplore.com/news/2023-07-generative-ai-scientists-artificial-automated.html