dialog context c:前k-1个utterances，conversational floor（1或0）,meta features(topic)
latent variable z: capture distribution of valid responses
x: response utterance
y:linguistic features(knowledge-guided CVAE)
p(z|c):prior network
p(x|z,c): response decoder，用q(x|z,c) recognition network来模拟
生成过程：sample z, generate x如图c所示
训练过程，如图b所示，通过max L目标函数得到q p的两个参数，从而得到z的分布。再由z生成response
The neural network architectures for the baseline and the proposed CVAE/kgCVAE models. The dashed blue connections only appear in kgCVAE
