Rethinking the Training Paradigm of Discrete Token-Based Multimodal LLMs: An Analysis of Text-Centric Bias.

@inproceedings{jo-2025-rethinking, author = {Wansik Jo and Jooyeong Na and Soyeon Hong and Seungtaek Choi and Hyunsouk Cho}, editor = {Meeyoung Cha and Carl Yang and Senjuti Basu Roy and Chanyoung Park and Zhenhui Jessie Li and Noseong Park and Carl Yang and Senjuti Basu Roy and Zhenhui Jessie Li and Jaap Kamps and Kijung Shin and Bryan Hooi and Lifang He}, title = {{Rethinking the Training Paradigm of Discrete Token-Based Multimodal LLMs: An Analysis of Text-Centric Bias}}, booktitle = {{Proceedings of the 34th ACM International Conference on Information and Knowledge Management, CIKM 2025, Seoul, Republic of Korea, November 10-14, 2025}}, pages = {1249--1259}, publisher = {ACM}, year = {2025}, url = {https://doi.org/10.1145/3746252.3761264}, doi = {https://doi.org/10.1145/3746252.3761264} }