model_in_dqn Find the explicit model in deep q network Exp. 1 Supervised Maze (identical as VIN) Current Results: 8x8: 99.02% (k=1,SAME/VALID,Batch_decay) 16x16: 97.18% (k=1,mpl_iter) 28x28: 95.07% (k=1,mpl_iter)