pith:3GM4RFGA
How to Guide Your Flow: Few-Step Alignment via Flow Map Reward Guidance
Reformulating guidance as deterministic optimal control lets the flow map steer trajectories to rewards in a single pass.
arxiv:2604.27147 v2 · 2026-04-29 · cs.LG · cs.AI
Add to your LaTeX paper
\usepackage{pith}
\pithnumber{3GM4RFGAGFGAC5TWUXX5BHCWO4}
Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge
Record completeness
Claims
FMRG matches or surpasses baselines across inverse problems, style transfer, human preferences, and VLM rewards with as few as 3 NFEs, giving at least an order-of-magnitude speedup in comparison to prior state of the art.
The deterministic optimal control reformulation of the guidance problem remains accurate and useful when discretized to very few steps and when the flow map is applied directly without additional approximation error or training.
FMRG is a training-free, single-trajectory guidance method for flow models derived from optimal control that achieves strong reward alignment with only 3 NFEs.
Cited by
Receipt and verification
| First computed | 2026-05-21T01:04:26.515890Z |
|---|---|
| Builder | pith-number-builder-2026-05-17-v1 |
| Signature | Pith Ed25519
(pith-v1-2026-05) · public key |
| Schema | pith-number/v1.0 |
Canonical hash
d999c894c0314c017676a5efd09c56770643adcf68f4025a59e700f8b8a70935
Aliases
· · · · ·Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/3GM4RFGAGFGAC5TWUXX5BHCWO4 \
| jq -c '.canonical_record' \
| python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: d999c894c0314c017676a5efd09c56770643adcf68f4025a59e700f8b8a70935
Canonical record JSON
{
"metadata": {
"abstract_canon_sha256": "e5d90b16e5401c58e65b5807a6d45f1d2793580e11b4b4859aab0e6d9bceeefb",
"cross_cats_sorted": [
"cs.AI"
],
"license": "http://creativecommons.org/licenses/by/4.0/",
"primary_cat": "cs.LG",
"submitted_at": "2026-04-29T19:56:53Z",
"title_canon_sha256": "3e507dbcf04cafc1041733300149f26b66da165b267a75b89b379adfd41b4525"
},
"schema_version": "1.0",
"source": {
"id": "2604.27147",
"kind": "arxiv",
"version": 2
}
}