L2Dir: Integrating L_2-Norm and Directional Alignment for Unsupervised Contrastive Representation Learning in Multimodal Retrieval

Multimodal representation learning primarily relies on contrastive objectives such as In-foNCE to align diverse modalities. However, these methods focus almost exclusively on directional alignment and often neglect the intrinsic role of embedding magnitudes (L 2 -norm) in the contrastive process. To bridge this gap, we propose L2Dir, a plug-and-play framework designed to optimize L 2 -norm alignment and Directional consistency jointly. As a highly efficient solution, L2Dir doesn't require extra data, distillation, or external supervision. It can be integrated seamlessly into existing pipelines by employing a lightweight MLP to reconstruct magnitudes from backbone features. Extensive evaluations across 95 tasks using UniIR and VLM2Vec-V2 frameworks demonstrate that L2Dir yields consistent and significant performance gains over established baselines across various backbones and scales, proving that explicit magnitude modeling is a versatile and potent method for refining unsupervised multimodal representations. The source code is available at https://github. com/tianyuzong/L2Dir_ACL2026 .