The paper investigates the expressive power of multimodal contrastive learning architectures by treating them as parameterized families of joint density estimators. It shows that the classic two‑tower CLIP model is a universal approximator for two modalities, while a common extension that sums pairwise similarities fails to approximate arbitrary joint distributions when three or more modalities are involved, though it can match all pairwise conditionals. To address this limitation, the authors introduce Hadamard‑CLIP, which adds a single learned weight vector to restore universal approximation for any number of modalities while retaining CLIP’s efficient retrieval capabilities.
By Andrew Stuart, Florian Wolf
arXiv:2606. 04180v1 Announce Type: new Abstract: Vision-language foundation models such as CLIP and SigLIP provide widely used representations for multimodal learning systems.
By Youqi Wu, Mohammad Jalali, Farzan Farnia
arXiv:2602. 07026v3 Announce Type: replace-cross Abstract: Despite the success of multimodal contrastive learning in aligning visual and linguistic representations, a persistent geometric anomaly, the Modality Gap, remains: embeddings of distinct modalities expressing identical semantics occupy systematically offset regions.
By Xiaomin Yu, Yi Xin, Yuhui Zhang, Wenjie Zhang, Chonghan Liu, Hanzhen Zhao, Chen Liu, Xiaoxing Hu, Ziyue Qiao, Hao Tang, Xiaobin Hu, Chengwei Qin, Hui Xiong, Yu Qiao, Shuicheng Yan
arXiv:2603. 02237v2 Announce Type: replace-cross Abstract: Representation steering offers a lightweight mechanism for controlling the behavior of large language models (LLMs) by intervening on internal activations at inference time.
By Laziz U. Abdullaev, Noelle Y. L. Wong, Ryan T. Z. Lee, Shiqi Jiang, Khoi N. M. Nguyen, Tan M. Nguyen
arXiv:2606. 13288v1 Announce Type: cross Abstract: Contrastively trained vision-language models like CLIP, have made remarkable progress in learning joint image-text representations, but still face challenges in compositional understanding.
By Wei Li, Zhen Huang, Xinmei Tian
arXiv:2605. 13352v2 Announce Type: replace Abstract: Standard dual-encoder vision-language models that map images and text to deterministic points on a shared unit hypersphere through $\ell_2$ normalization typically expose neither \emph{aleatoric} uncertainty (cross-modal ambiguity) nor \emph{epistemic} uncertainty (lack of training-distribution support).
By Mayank Nautiyal, Li Ju, Andreas Hellander, Ekta Vats, Prashant Singh