Calibrating Multimodal Learning

Huan Ma, Qingyang Zhang, Changqing Zhang, Bingzhe Wu, Huazhu Fu, Joey Tianyi Zhou, Qinghua Hu
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:23429-23450, 2023.

Abstract

Multimodal machine learning has achieved remarkable progress in a wide range of scenarios. However, the reliability of multimodal learning remains largely unexplored. In this paper, through extensive empirical studies, we identify current multimodal classification methods suffer from unreliable predictive confidence that tend to rely on partial modalities when estimating confidence. Specifically, we find that the confidence estimated by current models could even increase when some modalities are corrupted. To address the issue, we introduce an intuitive principle for multimodal learning, i.e., the confidence should not increase when one modality is removed. Accordingly, we propose a novel regularization technique, i.e., Calibrating Multimodal Learning (CML) regularization, to calibrate the predictive confidence of previous methods. This technique could be flexibly equipped by existing models and improve the performance in terms of confidence calibration, classification accuracy, and model robustness.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-ma23i, title = {Calibrating Multimodal Learning}, author = {Ma, Huan and Zhang, Qingyang and Zhang, Changqing and Wu, Bingzhe and Fu, Huazhu and Zhou, Joey Tianyi and Hu, Qinghua}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {23429--23450}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/ma23i/ma23i.pdf}, url = {https://proceedings.mlr.press/v202/ma23i.html}, abstract = {Multimodal machine learning has achieved remarkable progress in a wide range of scenarios. However, the reliability of multimodal learning remains largely unexplored. In this paper, through extensive empirical studies, we identify current multimodal classification methods suffer from unreliable predictive confidence that tend to rely on partial modalities when estimating confidence. Specifically, we find that the confidence estimated by current models could even increase when some modalities are corrupted. To address the issue, we introduce an intuitive principle for multimodal learning, i.e., the confidence should not increase when one modality is removed. Accordingly, we propose a novel regularization technique, i.e., Calibrating Multimodal Learning (CML) regularization, to calibrate the predictive confidence of previous methods. This technique could be flexibly equipped by existing models and improve the performance in terms of confidence calibration, classification accuracy, and model robustness.} }
Endnote
%0 Conference Paper %T Calibrating Multimodal Learning %A Huan Ma %A Qingyang Zhang %A Changqing Zhang %A Bingzhe Wu %A Huazhu Fu %A Joey Tianyi Zhou %A Qinghua Hu %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-ma23i %I PMLR %P 23429--23450 %U https://proceedings.mlr.press/v202/ma23i.html %V 202 %X Multimodal machine learning has achieved remarkable progress in a wide range of scenarios. However, the reliability of multimodal learning remains largely unexplored. In this paper, through extensive empirical studies, we identify current multimodal classification methods suffer from unreliable predictive confidence that tend to rely on partial modalities when estimating confidence. Specifically, we find that the confidence estimated by current models could even increase when some modalities are corrupted. To address the issue, we introduce an intuitive principle for multimodal learning, i.e., the confidence should not increase when one modality is removed. Accordingly, we propose a novel regularization technique, i.e., Calibrating Multimodal Learning (CML) regularization, to calibrate the predictive confidence of previous methods. This technique could be flexibly equipped by existing models and improve the performance in terms of confidence calibration, classification accuracy, and model robustness.
APA
Ma, H., Zhang, Q., Zhang, C., Wu, B., Fu, H., Zhou, J.T. & Hu, Q.. (2023). Calibrating Multimodal Learning. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:23429-23450 Available from https://proceedings.mlr.press/v202/ma23i.html.

Related Material