Follow the Mean: Reference-Guided Flow Matching
For practitioners of generative AI, this provides a simple, training-free method to control pretrained flow matching models using reference examples, avoiding costly fine-tuning or auxiliary networks.
The paper shows that flow matching models can be controlled by shifting the conditional endpoint mean, enabling training-free or amortized adaptation via reference examples. Reference-Mean Guidance controls color, identity, style, and structure in a frozen 4B model without fine-tuning, while Semi-Parametric Guidance matches unconditional DiT-B/4 quality on AFHQv2 with swappable reference sets.
Existing approaches to controllable generation typically rely on fine-tuning, auxiliary networks, or test-time search. We show that flow matching admits a different control interface: adaptation through examples. For deterministic interpolants, the velocity field is solely governed by a conditional endpoint mean; shifting this mean shifts the flow itself. This yields a simple principle for controllable generation: steer a pretrained model by changing the reference set it follows. We instantiate this idea in two forms. Reference-Mean Guidance is training-free: it computes a closed-form endpoint-mean correction from a reference bank and applies it to a frozen FLUX.2-klein (4B) model, enabling control of color, identity, style, and structure while keeping the prompt, seed, and weights fixed. Semi-Parametric Guidance amortizes the same idea through an explicit mean anchor and learned residual refiner, matching unconditional DiT-B/4 quality on AFHQv2 while allowing the reference set to be swapped at inference time. These results point to a broader direction: generative models that adapt through data, not parameter updates.