Probabilistic Embeddings for Frozen Vision-Language Models: Uncertainty Quantification with Gaussian Process Latent Variable Models

Aishwarya Venkataramanan, Paul Bodesheim, Joachim Denzler
Proceedings of the Forty-first Conference on Uncertainty in Artificial Intelligence, PMLR 286:4309-4328, 2025.

Abstract

Vision-Language Models (VLMs) learn joint representations by mapping images and text into a shared latent space. However, recent research highlights that deterministic embeddings from standard VLMs often struggle to capture the uncertainties arising from the ambiguities in visual and textual descriptions and the multiple possible correspondences between images and texts. Existing approaches tackle this by learning probabilistic embeddings during VLM training, which demands large datasets and does not leverage the powerful representations already learned by large-scale VLMs like CLIP. In this paper, we propose GroVE, a post-hoc approach to obtaining probabilistic embeddings from frozen VLMs. GroVE builds on Gaussian Process Latent Variable Model (GPLVM) to learn a shared low-dimensional latent space where image and text inputs are mapped to a unified representation, optimized through single-modal embedding reconstruction and cross-modal alignment objectives. Once trained, the Gaussian Process model generates uncertainty-aware probabilistic embeddings. Evaluation shows that GroVE achieves state-of-the-art uncertainty calibration across multiple downstream tasks, including cross-modal retrieval, visual question answering, and active learning.

Cite this Paper

BibTeX
@InProceedings{pmlr-v286-venkataramanan25a, title = {Probabilistic Embeddings for Frozen Vision-Language Models: Uncertainty Quantification with Gaussian Process Latent Variable Models}, author = {Venkataramanan, Aishwarya and Bodesheim, Paul and Denzler, Joachim}, booktitle = {Proceedings of the Forty-first Conference on Uncertainty in Artificial Intelligence}, pages = {4309--4328}, year = {2025}, editor = {Chiappa, Silvia and Magliacane, Sara}, volume = {286}, series = {Proceedings of Machine Learning Research}, month = {21--25 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v286/main/assets/venkataramanan25a/venkataramanan25a.pdf}, url = {https://proceedings.mlr.press/v286/venkataramanan25a.html}, abstract = {Vision-Language Models (VLMs) learn joint representations by mapping images and text into a shared latent space. However, recent research highlights that deterministic embeddings from standard VLMs often struggle to capture the uncertainties arising from the ambiguities in visual and textual descriptions and the multiple possible correspondences between images and texts. Existing approaches tackle this by learning probabilistic embeddings during VLM training, which demands large datasets and does not leverage the powerful representations already learned by large-scale VLMs like CLIP. In this paper, we propose GroVE, a post-hoc approach to obtaining probabilistic embeddings from frozen VLMs. GroVE builds on Gaussian Process Latent Variable Model (GPLVM) to learn a shared low-dimensional latent space where image and text inputs are mapped to a unified representation, optimized through single-modal embedding reconstruction and cross-modal alignment objectives. Once trained, the Gaussian Process model generates uncertainty-aware probabilistic embeddings. Evaluation shows that GroVE achieves state-of-the-art uncertainty calibration across multiple downstream tasks, including cross-modal retrieval, visual question answering, and active learning.} }
Endnote
%0 Conference Paper %T Probabilistic Embeddings for Frozen Vision-Language Models: Uncertainty Quantification with Gaussian Process Latent Variable Models %A Aishwarya Venkataramanan %A Paul Bodesheim %A Joachim Denzler %B Proceedings of the Forty-first Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2025 %E Silvia Chiappa %E Sara Magliacane %F pmlr-v286-venkataramanan25a %I PMLR %P 4309--4328 %U https://proceedings.mlr.press/v286/venkataramanan25a.html %V 286 %X Vision-Language Models (VLMs) learn joint representations by mapping images and text into a shared latent space. However, recent research highlights that deterministic embeddings from standard VLMs often struggle to capture the uncertainties arising from the ambiguities in visual and textual descriptions and the multiple possible correspondences between images and texts. Existing approaches tackle this by learning probabilistic embeddings during VLM training, which demands large datasets and does not leverage the powerful representations already learned by large-scale VLMs like CLIP. In this paper, we propose GroVE, a post-hoc approach to obtaining probabilistic embeddings from frozen VLMs. GroVE builds on Gaussian Process Latent Variable Model (GPLVM) to learn a shared low-dimensional latent space where image and text inputs are mapped to a unified representation, optimized through single-modal embedding reconstruction and cross-modal alignment objectives. Once trained, the Gaussian Process model generates uncertainty-aware probabilistic embeddings. Evaluation shows that GroVE achieves state-of-the-art uncertainty calibration across multiple downstream tasks, including cross-modal retrieval, visual question answering, and active learning.
APA
Venkataramanan, A., Bodesheim, P. & Denzler, J.. (2025). Probabilistic Embeddings for Frozen Vision-Language Models: Uncertainty Quantification with Gaussian Process Latent Variable Models. Proceedings of the Forty-first Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research 286:4309-4328 Available from https://proceedings.mlr.press/v286/venkataramanan25a.html.

Related Material