TY  - RPRT
TI  - Improving Image Captioning by Mimicking Human Reformulation Feedback at Inference-time
AU  - Uri Berger
AU  - Omri Abend
AU  - Lea Frermann
AU  - Gabriel Stanovsky
PY  - 2025
UR  - https://arxiv.org/abs/2501.04513
ID  - 2501.04513
ER  -