DFM: Interpolant-free Dual Flow Matching
This work addresses a computational bottleneck in flow-based generative models for researchers and practitioners, offering an incremental improvement over existing flow matching methods.
The paper tackles the computational expense of training continuous normalizing flows (CNFs) by proposing an interpolant-free dual flow matching (DFM) approach that avoids explicit vector field assumptions and optimizes both forward and reverse fields with a novel objective, achieving state-of-the-art performance in SMAP unsupervised anomaly detection.
Continuous normalizing flows (CNFs) can model data distributions with expressive infinite-length architectures. But this modeling involves computationally expensive process of solving an ordinary differential equation (ODE) during maximum likelihood training. Recently proposed flow matching (FM) framework allows to substantially simplify the training phase using a regression objective with the interpolated forward vector field. In this paper, we propose an interpolant-free dual flow matching (DFM) approach without explicit assumptions about the modeled vector field. DFM optimizes the forward and, additionally, a reverse vector field model using a novel objective that facilitates bijectivity of the forward and reverse transformations. Our experiments with the SMAP unsupervised anomaly detection show advantages of DFM when compared to the CNF trained with either maximum likelihood or FM objectives with the state-of-the-art performance metrics.