作者位次 #5/12ACL 2026Poster· Natural Language Processing
L 2 D ir: Integrating L\_2 -Norm and Directional Alignment for Unsupervised Contrastive Representation Learning in Multimodal Retrieval
Tianyu Zong · Rui Dai · Hongzhu Yi · Yuanxiang Wang · Zhenghao Zhang · Zhenyu Guan · Yujia Yang · Bingkang Shi · Yueyang Ding · Xiangxiang Chu · Kaikui Liu · Jungang Xu
摘要
Multimodal representation learning primarily relies on contrastive objectives such as InfoNCE to align diverse modalities. However, these methods focus almost exclusively on directional alignment and often neglect the intrinsic role of embedding magnitudes (L2-norm) in the contrastive process. To...