Pith. sign in

REVIEW 6 cited by

Source-Guided Flow Matching

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2508.14807 v2 pith:GRNN257Z submitted 2025-08-20 cs.LG

Source-Guided Flow Matching

classification cs.LG
keywords distributionfieldflowguidancevectorframeworkmatchingproblem
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Guidance of generative models is typically achieved by modifying the probability flow vector field through the addition of a guidance field. In this paper, we instead propose the Source-Guided Flow Matching (SGFM) framework, which modifies the source distribution directly while keeping the pre-trained vector field intact. This reduces the guidance problem to a well-defined problem of sampling from the source distribution. We theoretically show that SGFM recovers the desired target distribution exactly. Furthermore, we provide bounds on the Wasserstein error for the generated distribution when using an approximate sampler of the source distribution and an approximate vector field. The key benefit of our approach is that it allows the user to flexibly choose the sampling method depending on their specific problem. To illustrate this, we systematically compare different sampling methods and discuss conditions for asymptotically exact guidance. Moreover, our framework integrates well with optimal flow matching models since the straight transport map generated by the vector field is preserved. Experimental results on synthetic 2D benchmarks, physics-informed generative tasks, and imaging inverse problems demonstrate the effectiveness and flexibility of the proposed framework.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 6 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. The Right Measure for Physics-Constrained Generation: A Co-Area Correction for Posterior-Consistent PDE Inverse Problems

    cs.LG 2026-06 unverdicted novelty 7.0

    Physics-constrained generative sampling for PDE inverse problems omits a co-area Jacobian factor required for correct posterior sampling; CoCoS corrects it to match the true distribution.

  2. Conflict-Aware Additive Guidance for Flow Models under Compositional Rewards

    cs.AI 2026-05 unverdicted novelty 6.0

    Conflict-Aware Additive Guidance (g^car) is a lightweight learnable method that dynamically resolves gradient conflicts to prevent off-manifold drift in compositional guided sampling for flow models.

  3. Na\"ive PAINE: Lightweight Text-to-Image Generation Improvement with Prompt Evaluation

    cs.CV 2026-03 conditional novelty 6.0

    A lightweight predictor ranks initial noises by expected human-preference score for a prompt, selecting the best few for diffusion generation and reporting prompt difficulty.

  4. Flow Matching Calibration for Simulation-Based Inference under Model Misspecification

    stat.ML 2025-09 conditional novelty 6.0

    FMCPE trains a flow-matching correction that transports samples from a simulation-based posterior estimator toward the true posterior, using only tens to hundreds of ground-truth calibration pairs.

  5. DS@GT ARC at ImageCLEFmed GANs 2026: Geometric Filtering for Privacy-Preserving CT Slice Generation

    cs.CV 2026-07 conditional novelty 5.0

    Geometric filtering of synthetic CT slices reduces nearest-neighbor and membership-inference leakage, but patient re-identification leakage remains near 0.97–1.0 across all submissions.

  6. Can We Predict The Human Preference For Text-to-Image Content Prior To Generation And Is It Even Useful To Do So?

    cs.CV 2026-06 unverdicted novelty 3.0

    Exploration of pre-generation prediction of human preference metrics (HPM) from noise seeds in diffusion models to improve output quality with negligible added cost.