We aim to showcase that reinforcement learning (RL) or machine learning (ML) with GPUs delivers the best benchmark performance for large-scale nonconvex and NP-complete problems ... and Jianliang Xu. ...