SLIP: Learning Transferable Sensor Models via Language-Informed Pretraining
Yuliang Chen, Arvind Pillai, Yu Yvonne Wu, Nicholas C. Jacobson, Andrew Campbell
arXiv 2603.11950· 2026
IMUPPGECGEEGFoundation ModelOpen Weights
Summary
An open-weights (MIT) sensor-language FM that is agnostic to channel count and sampling rate. That suits Clarity V2's mixed-rate PPG (green/red/IR) + IMU streams without retraining the patch embedder. It includes PPG evaluation tasks (PPG-BP-derived) and a SLIP-1B variant is one of the Lapras backbones.
Key facts
- Form Factor
- other
- License
- Not stated
Updated: Oct 12, 2026
Related Research
- A lightweight ECG model with teacher-derived integrated medical semantic preservation for cardiac disease screening (ECGFM-U-Lite)
Compresses a unified multitask ECG foundation model (rhythm/morphology + cardiac structure-function
- A robust PPG foundation model using multimodal physiological supervision
Pretrains a PPG-only foundation model (1D ResNet-26 encoder) on MIMIC-III Waveform Database Matched
- AnyPPG: An ECG-Guided PPG Foundation Model Trained on Over 100,000 Hours of Recordings for Holistic Health Profiling
AnyPPG pretrains a PPG encoder (Net1D/ResNet-style, 512-dim) using a CLIP-style cross-modal contrast