Prof. Dr. Heiko Weber from the FAU Profile Center Light.Matter.QuantumTechnologies is responsible for the sustainable handling of data gained in research. In our interview, he talks about the potential of big data in physics.
The sustainable handling of research data is a major topic at the Profile Center “Light.Material.QuantumTechnologies.” What data do you collect?
It can be terabytes of data per day, for example the exact moment photons hit a detector, which are then used to compare with measurements taken from a second detector. Or it may be manageable tables of voltages, where measurements were gathered over one hour for each data point. In our work in the area of solid-state physics and optics, we usually carry out research in small teams using innovative technology to investigate any number of new research questions, leading to great variety in the data we handle. By collecting data systematically, sustainably and in a structured manner, we hope to be able to recognise brand new connections that have so far remained hidden to individual researchers.
What does “sustainable” mean in the context of data handling? What are the benefits, in your opinion?
Sustainable means that researchers who have not been involved in the research or computer systems that use artificial intelligence will be able to understand our experiments down to the smallest detail even several years from now, and will be able to learn from them. It also means that data can be compared to other experiments conducted in another area and possibly considering another question. This “interoperability” is our overriding goal, and it will entail a huge added value for research.
How does the Profile Center intend to make its data more sustainable?
In Germany, we are in the lucky position that we have a widespread national research data infrastructure initiative (NFDI initiative) that is doing conceptual and technical groundwork. In particular, we make an extremely active contribution to the NFDI consortium for solid-state physics, FAIRmat. At the Profile Center we are also working on transferring these concepts and software solutions to device physics, quantum technology and optics, where no conventions have been established to date. We aim to define consistent data structures in the mentioned pioneering fields. It is even slightly easier to establish new structures there than in established areas.
University teaching is another important aspect: At FAU, we have already done pioneering work to teach students to use electronic lab books and to teach them skills in Python programming and comprehensive data skills in general from an early stage. We believe that this will be an important lever when it comes to actively shaping the transition to data-driven science.
Why is this effort worthwhile?
For centuries now, science has thrived on the interplay between hard, detailed work and the breakthroughs it enables. Breakthroughs in data-driven science will rely on comprehensive and well-structured datasets. With our background in research data management and our ambitious approach of breaking new ground in previously unexplored research areas, the Profile Center Light.Material.QuantumTechnologies is the right place to establish future standards.
Do you use data from other teams of researchers in your work?
I have often consulted data from other groups in order to gain a better understanding of apparent contradictions, but this line of questioning tends to come to a dead-end due to a lack of documentation. However, since we have started to pursue a big data approach in our own experiments, and even though it is still on a rather modest scale, we have already been able to discover fascinating principles that have previously remained hidden. It clearly works!
Prof. Dr. Heiko Weber from the FAU Profile Center Light.Matter.QuantumTechnologies is responsible for the sustainable handling of data gained in research. In our interview, he talks about the potential of big data in physics.
The sustainable handling of research data is a major topic at the Profile Center “Light.Material.QuantumTechnologies.” What data do you collect?
It can be terabytes of data per day, for example the exact moment photons hit a detector, which are then used to compare with measurements taken from a second detector. Or it may be manageable tables of voltages, where measurements were gathered over one hour for each data point. In our work in the area of solid-state physics and optics, we usually carry out research in small teams using innovative technology to investigate any number of new research questions, leading to great variety in the data we handle. By collecting data systematically, sustainably and in a structured manner, we hope to be able to recognise brand new connections that have so far remained hidden to individual researchers.
What does “sustainable” mean in the context of data handling? What are the benefits, in your opinion?
Sustainable means that researchers who have not been involved in the research or computer systems that use artificial intelligence will be able to understand our experiments down to the smallest detail even several years from now, and will be able to learn from them. It also means that data can be compared to other experiments conducted in another area and possibly considering another question. This “interoperability” is our overriding goal, and it will entail a huge added value for research.
How does the Profile Center intend to make its data more sustainable?
In Germany, we are in the lucky position that we have a widespread national research data infrastructure initiative (NFDI initiative) that is doing conceptual and technical groundwork. In particular, we make an extremely active contribution to the NFDI consortium for solid-state physics, FAIRmat. At the Profile Center we are also working on transferring these concepts and software solutions to device physics, quantum technology and optics, where no conventions have been established to date. We aim to define consistent data structures in the mentioned pioneering fields. It is even slightly easier to establish new structures there than in established areas.
University teaching is another important aspect: At FAU, we have already done pioneering work to teach students to use electronic lab books and to teach them skills in Python programming and comprehensive data skills in general from an early stage. We believe that this will be an important lever when it comes to actively shaping the transition to data-driven science.
Why is this effort worthwhile?
For centuries now, science has thrived on the interplay between hard, detailed work and the breakthroughs it enables. Breakthroughs in data-driven science will rely on comprehensive and well-structured datasets. With our background in research data management and our ambitious approach of breaking new ground in previously unexplored research areas, the Profile Center Light.Material.QuantumTechnologies is the right place to establish future standards.
Do you use data from other teams of researchers in your work?
I have often consulted data from other groups in order to gain a better understanding of apparent contradictions, but this line of questioning tends to come to a dead-end due to a lack of documentation. However, since we have started to pursue a big data approach in our own experiments, and even though it is still on a rather modest scale, we have already been able to discover fascinating principles that have previously remained hidden. It clearly works!