Deep Q-Learning for Directed Acyclic Graph Generation
This addresses a domain-specific problem for applications requiring DAGs, but it appears incremental as it adapts existing deep reinforcement learning to a graph generation task.
The paper tackled the problem of generating directed acyclic graphs (DAGs) with specified structures, which is challenging as most methods produce undirected graphs, and demonstrated that their deep Q-learning method can generate DAGs meeting criteria in sparse reward environments.
We present a method to generate directed acyclic graphs (DAGs) using deep reinforcement learning, specifically deep Q-learning. Generating graphs with specified structures is an important and challenging task in various application fields, however most current graph generation methods produce graphs with undirected edges. We demonstrate that this method is capable of generating DAGs with topology and node types satisfying specified criteria in highly sparse reward environments.