A Deep Reinforcement Learning Framework with Formal Verification

Artificial Intelligence (AI) and data are reshaping organizations businesses. Human Resources (HR) management talent development make no exception, as they tend to involve more automation growing quantities of data. Because this brings implications on workforce, career transparency, equal opportunities, overseeing what fuels AI analytical models, their quality standards, integrity, correctness becomes an imperative for those aspiring such systems. Based ontology transformation B-machines, article presents approach constructing a valid error-free agent with Deep Reinforcement Learning (DRL). In short, the agent's policy is built framework we called Multi State-Actor (MuStAc) using decentralized training approach. Its purpose predict both relevant steps employees, based profiles company pathways (observations). Observations can comprise various elements current occupation, past experiences, performance, skills, qualifications, so on. The takes in all these observations outputs next recommended step, environment set combination HR Event-B model, which generates action spaces respect formal properties. model properties derived OWL B transformation.