News Release 

Can artificial intelligence lead scientific discoveries?

University of Konstanz

Can machines be agents with the capacity for autonomous action? Can they be creative and produce something genuinely new? Philosopher Professor Thomas Müller from the University of Konstanz and physicist Professor Hans Briegel from the University of Innsbruck receive a total of 825,000 euros over four years from the Volkswagen Foundation through the funding initiative "Off the Beaten Track" to explore the role of artificial intelligence (AI) in basic research.

The aim is to provide a conceptual framework for the description, evaluation and regulation of the current rapid technological development surrounding the use of AI - a development that affects the field of economy as well as the arts. The researchers will focus on clarifying the concepts of agency, creativity and authorship in research.

Artificial intelligence and the concept of agency

The funded project "The future of creativity in basic research: Can artificial agents be authors of scientific discoveries?" combines methods and techniques from the fields of philosophy, physics and artificial intelligence (AI) in an interdisciplinary approach. Theoretical philosopher Müller and theoretical physicist Briegel, who has been holding a visiting professorship at the University of Konstanz since 2017, funded by the state of Baden-Württemberg, have been collaborating for almost ten years. Their shared interest revolves around the concept of agency - understood as the opposite of an uncontrolled bodily reaction - and its application in the natural sciences. They believe this clarification to be a matter of increasing urgency in view of the developments in the natural sciences in the 21st century. In research experiments, for example, more and more work steps that were previously performed by humans are being outsourced to AI systems. Hence the question: Can AI systems be self-acting agents that make their own scientific discoveries?

When a new Jupiter moon is discovered with the help of a telescope, the role of the telescope as an instrument is undisputed: it serves to expand the human senses. But what happens when AI programmes are involved, such as so-called Generative Adversarial Networks (GAN), a special combination of artificial neural networks, or the learning method Projective Simulation developed by Hans Briegel? The physicist and his colleagues have been using Projective Simulation, for example, to design experiments that can produce novel quantum states in the laboratory. In this context, the AI, which is based on the methods of reinforcement learning, has advanced itself based on the experiences made in the process. By gathering feedback on successful setups, the system has gradually learned which combination of experimental elements would frequently result in the generation of such states. Hans Briegel: "We may not be far from an AI proposing experiments that will lead to the discovery of new quantum phenomena. This will then also raise the question what part we or the AI respectively have in such a discovery."

Focus areas of the project

The project is divided into three focus areas: The role of experiments in AI-driven research, specific case studies on the implementation of AI methods in basic research in physics, and the question of authorship in AI-driven research. Thomas Müller says on the subject of authorship: "We don't know if in ten years an AI will be named as the author on a research paper. What we are interested in is the underlying philosophical and conceptual question: At what point would we attribute to a system the ability, literally, to be curious, to be surprised, to intervene, to have a new idea?"

Collaboration partners

The two researchers are joined by three doctoral candidates who will work on the individual focus areas. The team will work in close exchange with working groups in whose research machine learning and artificial intelligence play an increasing role. At the University of Konstanz, research groups from the Cluster of Excellence "Centre for the Advanced Study of Collective Behaviour" will be involved as collaboration partners. Further collaborations are planned in the field of climate research with Professor Tapio Schneider at the California Institute of Technology (USA), who is working on a new climate model as part of the CliMA project. In the field of quantum optics, collaboration partners include, among others, the working group of Professor Markus Aspelmeyer at the Institute for Quantum Optics and Quantum Information of the Austrian Academy of Sciences (IQOQI) in Innsbruck and Vienna, who are developing micromechanical experiments for the measurement of ultra-weak gravitational fields.

###

Key facts:

    - The project "The future of creativity in basic research: Can artificial agents be authors of scientific discoveries?"

    - Project partnership between philosopher Professor Thomas Müller from the University of Konstanz and physicist Professor Hans Briegel from the University of Innsbruck

    - Study on the role of artificial intelligence in basic research

    - Supported by the Volkswagen Foundation with 825,000 euros over four years

    - Collaborations with the Cluster of Excellence "Centre for the Advanced Study of Collective Behaviour" at the University of Konstanz, the CliMA project at the California Institute of Technology, and the Institute for Quantum Optics and Quantum Information at the Austrian Academy of Sciences.

Note to editors:

A photo is available for download here: https://cms.uni-konstanz.de/fileadmin/pi/fileserver/2020/Bilder/kuenstliche_intelligenz.jpg

Caption: Philosopher Professor Thomas Müller (right) and physicist Professor Hans Briegel receive funding from the Volkswagen Foundation for their study on the role of artificial intelligence (AI) in basic research.

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.