2 November 2023
Image and Video Processing
Proposes a dynamic multimodal information bottleneck framework for fusing data
Employs techniques from information theory to distill predictive signals
Demonstrates superior performance on COVID-19 mortality prediction
Significantly more robust to noise and missing data than other methods
Achieves state-of-the-art results on biomedical multimodal classification tasks
Distilling multimodal COVID-19 data for robust predictions
This paper proposes a deep learning approach to fuse multimodal medical data like images and clinical records. It introduces techniques to filter out redundant and noisy information, while retaining predictive signals. On COVID-19 and other biomedical datasets, the method achieves state-of-the-art performance and is highly robust to missing data.
No comments yet, be the first to start the conversation...
Sign up to comment on this paper