Device Passport: Enabling Spatio-Temporal Pretrained Models to Generalize Across Input Layouts
2026-06-30 • Machine Learning
Machine Learning
AI summaryⓘ
The authors looked at how to make biosignal models work well when new devices have different sensor layouts and there isn’t much data for those new designs. They focused on how to represent each sensor channel's information by combining what the channel measures and metadata about it. They created a method called Device Passport, which uses this combined information to better transfer knowledge from one layout to another. Their experiments show that Device Passport works better than previous methods when moving between different device layouts. This means carefully designing how channels are represented can help reuse big biosignal models across devices.
biosignal modelschannel embeddingdevice layouttransfer learningpretrainingmetadatafunctional activityEEGmixture modelsfoundation models
Authors
Geeling Chau, Ran Liu, Juri Minxha, Wenhui Cui, Erdrin Azemi, Ellen L. Zippi, Behrooz Mahasseni, Christopher M. Sandino
Abstract
New device layouts pose a challenging modeling problem due to the lack of large datasets for each specific layout. Biosignal foundation models offer a plausible solution if they are able to generalize to new layouts effectively. To improve cross-layout transfer, we study how different channel embedding techniques behave when pretraining layouts differ substantially from the downstream decoding layout. We propose Device Passport, a new channel embedding technique that learns experts and mixture models that take each channel's functional activity and metadata as input. This contrasts with prior embedding methods, which typically use only functional information or only metadata to look up learned or fixed positional embeddings. Across controlled subset-transfer experiments and realistic transfer to ear-EEG, Device Passport is competitive overall and improves over the strongest learned baseline in the layout-transfer regimes that motivate this work. These results suggest that channel embedding design is a key consideration when reusing large-scale pretrained biosignal models on new devices.