@misc{indiciae5ebc204c1421,
  title = {Switching to Discriminative Image Captioning by Relieving a Bottleneck of Reinforcement Learning},
  author = {Ukyo Honda and Taro Watanabe and Yuji Matsumoto},
  year = {2022},
  url = {https://arxiv.org/abs/2212.03230},
  note = {Source identifier: 2212.03230}
}