Fig. 1

Human-in-the-loop de novo molecular design: an AI-assistant helps a chemist to decide parameters of an MPO objective function \({S}_{r,t}\left(x\right)\) iteratively at round \(r\) and iteration \(t\), where \(r\) are rounds of goal-directed molecule generation with a de novo design tool, and \(t\) are online interactions with a chemist. The objective consists of K molecular properties \({c}_{k}\left(x\right)\) with relative weights \({w}_{k}\). The utility of the \(k\):th property is measured using a desirability function \({\phi }_{r,t,k}\) that defines the range of good property values. At each iteration, the method selects a molecule \({x}_{r,t}\) to query, which the chemist evaluates with feedback \(y\). The method then adapts \({S}_{r,t}\left(x\right)\) based on the feedback by estimating the parameters of \({\phi }_{r,t,k}\) to match the chemist’s underlying goal