High-Performance Self-Supervised Learning by Joint Training of Flow Matching
This addresses computational efficiency and performance bottlenecks for self-supervised learning in industrial and edge AI applications, particularly for wearable sensor data.
The paper tackles the trade-off between generative quality and discriminative performance in diffusion models for self-supervised learning by proposing FlowFM, which jointly trains a representation encoder and conditional flow matching generator. Results show FlowFM reduces training time by 50.4% compared to diffusion-based approaches, surpasses state-of-the-art SSL methods on all five datasets, and achieves up to 51.0x inference speedup while maintaining high generative quality.
Diffusion models can learn rich representations during data generation, showing potential for Self-Supervised Learning (SSL), but they face a trade-off between generative quality and discriminative performance. Their iterative sampling also incurs substantial computational and energy costs, hindering industrial and edge AI applications. To address these issues, we propose the Flow Matching-based Foundation Model (FlowFM), which jointly trains a representation encoder and a conditional flow matching generator. This decoupled design achieves both high-fidelity generation and effective recognition. By using flow matching to learn a simpler velocity field, FlowFM accelerates and stabilizes training, improving its efficiency for representation learning. Experiments on wearable sensor data show FlowFM reduces training time by 50.4\% compared to a diffusion-based approach. On downstream tasks, FlowFM surpassed the state-of-the-art SSL method (SSL-Wearables) on all five datasets while achieving up to a 51.0x inference speedup and maintaining high generative quality. The implementation code is available at https://github.com/Okita-Laboratory/jointOptimizationFlowMatching.