Synthesis Predictions in AI-Based Materials Design

Chemical research is primarily focused on discovering new molecules and materials that possess desired properties. To improve this process, an effective strategy involves utilizing all accessible knowledge and data to strategically design future materials using computers. However, a notable hurdle in digital discovery is that many molecules and materials designed computationally are often impractical to synthesize in the laboratory. In this presentation, I will delve into the synthesizability of molecules and materials, including the prediction of synthesis pathways (retrosynthesis) and chemical reactivity. Additionally, I will explore several challenges and opportunities that await further advancements in accelerated chemical platforms.