@misc{indiciae6918e6ab6920,
  title = {EBMs Trained with Maximum Likelihood are Generator Models Trained with a Self-adverserial Loss},
  author = {Zhisheng Xiao and Qing Yan and Yali Amit},
  year = {2021},
  url = {https://arxiv.org/abs/2102.11757},
  note = {Source identifier: 2102.11757}
}