This project performs exploratory research to policy learning methods based on deep neural networks with model-based approaches to heuristic search for domain-independent planning as studied in the AI Planning community. The project is supported by IBM Research with an IBM Faculty Award.