Numerous technological advancements are driven by materials development, in fields as diverse as energy, environmental protection, information technology, or medicine and health. As a result, an acceleration of materials development can make a strong contribution to solving major societal challenges. Recognizing that virtual material design itself is emerging as one of the grand challenges in the research field 'information' of the Helmholtz association, we have established a joint virtual materials design laboratory and are holding an international workshop on this topic under the CECAM umbrella in 2021.

In this follow-up flagship workshop, we aim to address present challenges in the design of complex high-technology products and to establish new information-based materials development strategies to fulfill the requirements of the economy of the 21st century. In order to achieve this goal, the traditional unidirectional process of product development (new materials –> new components –> new products) needs to be replaced by an inverse design of tailor-made new materials and device components based on specific system and device requirements.

As result of decades of method development and continuously increasing computational power, the possibility of a paradigm shift has emerged, in which the still indispensable experimental materials research driven by trial and error, is complemented by computer-based approaches to systematically design novel materials. In order to go beyond the current limitations of virtual materials design, we want to address the following critical issues:

• Concepts for the computational realization of digital twins

• Tackling the size of chemical spaces beyond exhaustive search

• Integrating experiment on the fly and autonomous discovery

• Machine learning applications beyond electronic structure

• The synthetic challenge: how can the proposed materials be made?

• Integration of processing challenges into materials design

• Adaptation of parameters in scale-bridging workflows in materials science

The proposed event will be organized in four half-day sessions. In session 1, we will review current method developments in multiscale modeling combined with scientific workflow technology. As traditional materials-discovery strategies are increasingly complemented by information-based methods, including deep learning and artificial intelligence, we will dedicate session 2 to the application of machine learning methods in materials design. In session 3, we want to discuss the development and use of digital twins of materials in order to not only digital represent devices and their components but also the constituting materials. Finally, session 4 will focus on examples of accelerated materials design where simulations have allowed for experimental breakthroughs in physics, chemistry, biotechnology and materials science.