This internship offers students the opportunity to participate in the development of explainable machine learning models (Explainable AI), which represent a rapidly developing field with the aim of making black-box models more accessible. These methods make it possible to reveal how models—such as neural networks—learn and arrive at their predictions. Special attention is paid to how to choose an appropriate “vocabulary” of terms to explain the results to people and how to use this vocabulary in the construction of the models themselves. An approach combining human interaction and machine learning has the potential to accelerate the development of scientific theories and streamline experimental work.
What will the internship involve?
- Introduction to the concepts of explainable artificial intelligence (Explainable AI)
- Exploring the limits of black-box models and the risks of misinterpretation
- Working on the design and testing of models with explainable principles
- Collaborating with a supervisor on research into the use of AI in plasma physics and astronomy
- Practical experience with the use of AI for the interactive development of scientific theories
What we need:
- Basic knowledge of machine learning or data analysis
- Willingness to learn new approaches and work at the intersection of AI and scientific research
- Interest in the application of AI in the natural sciences
What we offer:
- Guidance from an experienced supervisor with practical experience in the application of AI in physics and astronomy
- Involvement in current research on explainable machine learning models
- The opportunity to contribute to the development of a new generation of tools for scientific work
- The inspiring environment of ÚTIA AV ČR
If you are attracted to artificial intelligence, but also want the results to be understandable to people and help accelerate the discovery of new scientific theories, then this internship is the opportunity for you.