Skip to main content
. 2018 Apr 25;34(18):3169–3177. doi: 10.1093/bioinformatics/bty323

Fig. 6.

Fig. 6.

Performance evaluation of the deep reinforcement learning algorithm for cell movement modeling. (a) The cumulative rewards generally goes upward, but tends to be noisy. (b) The loss tends to oscillate because of the implementation of the experience replay and the target network. (c) The average action value grows smoothly over time