Researchers have developed a new AI system called SEISMO that can optimize molecules for specific properties in a more efficient way than current methods. This is particularly useful in the pharmaceutical industry where discovering new drugs is a time-consuming and costly process. SEISMO uses natural language processing to understand the desired properties of a molecule and conditions its proposals on a range of factors, including post-hoc explainability methods. This approach has been shown to improve sample efficiency across various tasks relevant to drug discovery. --- Why it matters: This matters because it could lead to faster and more cost-effective development of new drugs, which is crucial in the pharmaceutical industry where time-to-market is often a major bottleneck. Source: https://arxiv.org/abs/2602.00663