Speaker
Description
Flow matching has recently emerged as a powerful paradigm for generative modeling and has been extended to probabilistic forecasting of spatio-temporal dynamics. However, the impact of the specific choice of probability path model on forecasting performance remains under-explored. Motivated by dynamical optimal transport and the Schrödinger bridge perspective, we present a novel probability path model, together with the theoretical framework and efficient algorithms, for applying flow matching in latent space to improve forecasting performance. Our empirical results across various dynamical system benchmarks show that our model achieves faster convergence during training and improved predictive performance compared to existing probability path models. Importantly, our approach is efficient during inference, requiring only a few sampling steps. This makes our proposed model practical for real-world applications and opens new avenues for probabilistic forecasting.
