Our findings on multi-modal learning for brain signals are finally live ! EEG-CLIP aligns physiological patterns with text reports through contrastive learning, allowing recording retrieval and zero-shot classification

paper : www.frontiersin.org/journals/rob...
code : github.com/tidiane-cama...

2nd place at #CVPR2025 Interactive Segmentation Challenge!
At @uniklinik-fr.bsky.social, we're building vision models that adapt from user clicks/corrections - to better integrate expert knowledge into our segmentation tools
📖 Paper: openreview.net/pdf?id=EnSf2...
💻 Code: github.com/tidiane-cama...