Discriminative Flow Matching Via Local Generative Predictors

TL;DR

This paper introduces Discriminative Flow Matching, a novel framework that models classification and detection as a continuous transport process, integrating generative and discriminative learning for improved robustness and flexibility.

Contribution

It proposes a new method that learns vector fields for class and object detection tasks, enabling iterative refinement and compatibility with various neural architectures.

Findings

Effective in image classification tasks.

Extended successfully to object detection.

Provides robustness and flexibility in inference.

Abstract

Traditional discriminative computer vision relies predominantly on static projections, mapping input features to outputs in a single computational step. Although efficient, this paradigm lacks the iterative refinement and robustness inherent in biological vision and modern generative modelling. In this paper, we propose Discriminative Flow Matching, a framework that reformulates classification and object detection as a conditional transport process. By learning a vector field that continuously transports samples from a simple noise distribution toward a task-aligned target manifold -- such as class embeddings or bounding box coordinates -- we are at the interface between generative and discriminative learning. Our method attaches multiple independent flow predictors to a shared backbone. These predictors are trained using local flow matching objectives, where gradients are computed independently for each block. We formulate this approach for standard image classification and extend it to the complex task of object detection, where targets are high-dimensional and spatially distributed. This architecture provides the flexibility to update blocks either sequentially to minimise activation memory or in parallel to suit different hardware constraints. By aggregating the predictions from these independent flow predictors, our framework enables robust, generative-inspired inference across diverse architectures, including CNNs and vision transformers.