Well Begun is Half Done: Training-Free and Model-Agnostic Semantically Guaranteed User Representation Initialization for Multimodal Recommendation Conference: Proceedings of the 49th International ACM SIGIR Conference on Research and Development in Information Retrieval; July 20–24, 2026; Melbourne, VIC, Australia Proceedings of the 49th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR ’26), July 20–24, 2026, Melbourne, VIC, Australia DOI: 10.1145/3805712.3809623 ISBN: 979-8-4007-2599-9/2026/07 CCS: Information systems Recommender systems;
Jinfeng Xu, Zheyu Chen, Shuo Yang, Jinze Li, Hewei Wang, Jianheng Tang, Wei Wang, Xiping Hu, Edith C. H. Ngai · 2026
Read Well Begun is Half Done: Training-Free and Model-Agnostic Semantically Guaranteed User Representation Initialization for Multimodal Recommendation Conference: Proceedings of the 49th International ACM SIGIR Conference on Research and Development in Information Retrieval; July 20–24, 2026; Melbourne, VIC, Australia Proceedings of the 49th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR ’26), July 20–24, 2026, Melbourne, VIC, Australia DOI: 10.1145/3805712.3809623 ISBN: 979-8-4007-2599-9/2026/07 CCS: Information systems Recommender systems; by Jinfeng Xu, Zheyu Chen, Shuo Yang, Jinze Li, Hewei Wang, Jianheng Tang, Wei Wang, Xiping Hu, Edith C. H. Ngai (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-14839.txt.
Description
Xu and eight colleagues target a neglected gap in multimodal recommendation: items start modality-rich while users start from random vectors, leaving a semantic gap. Their SG-URInit builds each user's starting representation from modality features of touched items plus global cluster features. The result captures both item-level and cluster-level semantics. Training-free and model-agnostic, it slots into existing models with no extra training cost. Experiments on real datasets show clear performance gains, plus relief for item cold-start and faster convergence.
Catalog record
| CREATOR | Jinfeng Xu, Zheyu Chen, Shuo Yang, Jinze Li, Hewei Wang, Jianheng Tang, Wei Wang, Xiping Hu, Edith C. H. Ngai |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 45322 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.14839