Solving NP-Hard Problems on Graphs with Extended AlphaGo Zero
There have been increasing challenges to solve combinatorial optimization problems by machine learning. Khalil et al. proposed an end-to-end reinforcement learning framework, S2V-DQN, which...