DyCa-GRPO: Calibrated and Efficient Learning-to-Rank for Multimodal Retrieval with Human Feedback.

Ning Han|Xiubo Liang


Anthology ID:DBLP:conf/mir/HanL26
Volume:Proceedings of the 2026 International Conference on Multimedia Retrieval, ICMR 2026, Amsterdam, The Netherlands, June 16-19, 2026
Year:2026
Venue:International Conference on Multimedia Retrieval (ICMR)
Publisher:ACM
Pages:215-222
URL:https://doi.org/10.1145/3805622.3810627
DOI:https://doi.org/10.1145/3805622.3810627
DBLP:conf/mir/HanL26
BibTeX:
@inproceedings{han-2026-dycagrpo, author = {Ning Han and Xiubo Liang}, editor = {Stevan Rudinac and Zi Huang and Marcel Worring and Lucia Vadicamo and Xirong Li and Pascal Mettes and Lizi Liao}, title = {{DyCa-GRPO: Calibrated and Efficient Learning-to-Rank for Multimodal Retrieval with Human Feedback}}, booktitle = {{Proceedings of the 2026 International Conference on Multimedia Retrieval, ICMR 2026, Amsterdam, The Netherlands, June 16-19, 2026}}, pages = {215--222}, publisher = {ACM}, year = {2026}, url = {https://doi.org/10.1145/3805622.3810627}, doi = {https://doi.org/10.1145/3805622.3810627} }