RelCon: Relative Contrastive Learning for a Motion Foundation Model for Wearable Data
This work addresses the need for generalizable motion foundation models in wearable computing, though it appears incremental as it builds on existing contrastive learning and foundation model concepts.
The authors tackled the problem of learning general motion representations from wearable accelerometer data by introducing RelCon, a self-supervised relative contrastive learning approach, which achieved state-of-the-art performance on tasks like human activity recognition and gait metric regression using a model trained on 1 billion segments from 87,376 participants.
We present RelCon, a novel self-supervised Relative Contrastive learning approach for training a motion foundation model from wearable accelerometry sensors. First, a learnable distance measure is trained to capture motif similarity and domain-specific semantic information such as rotation invariance. Then, the learned distance provides a measurement of semantic similarity between a pair of accelerometry time-series, which we use to train our foundation model to model relative relationships across time and across subjects. The foundation model is trained on 1 billion segments from 87,376 participants, and achieves state-of-the-art performance across multiple downstream tasks, including human activity recognition and gait metric regression. To our knowledge, we are the first to show the generalizability of a foundation model with motion data from wearables across distinct evaluation tasks.