Multidimensional Adaptive Coefficient for Inference Trajectory Optimization in Flow and Diffusion

TL;DR

This paper introduces the Multidimensional Adaptive Coefficient (MAC), a plug-in module that enhances flow and diffusion models by enabling multidimensional and trajectory-wise adaptive inference, improving generative quality efficiently.

Contribution

The paper presents MAC, a novel module that extends unidimensional coefficients to multidimensional ones and allows inference trajectory adaptation, trained via simulation-based feedback.

Findings

MAC improves generative quality across frameworks.

Training with MAC is highly efficient.

MAC demonstrates versatility across datasets.

Abstract

Flow and diffusion models have demonstrated strong performance and training stability across various tasks but lack two critical properties of simulation-based methods: freedom of dimensionality and adaptability to different inference trajectories. To address this limitation, we propose the Multidimensional Adaptive Coefficient (MAC), a plug-in module for flow and diffusion models that extends conventional unidimensional coefficients to multidimensional ones and enables inference trajectory-wise adaptation. MAC is trained via simulation-based feedback through adversarial refinement. Empirical results across diverse frameworks and datasets demonstrate that MAC enhances generative quality with high training efficiency. Consequently, our work offers a new perspective on inference trajectory optimality, encouraging future research to move beyond vector field design and to leverage training-efficient, simulation-based optimization.