Pith Number

pith:KKLSRPH4

pith:2026:KKLSRPH4474HO5EJ6H5QDCT6FE

not attested not anchored not stored refs pending

World-R1: Reinforcing 3D Constraints for Text-to-Video Generation

Bohan Zhuang, Donny Y. Chen, Weijie Wang, Xiaoxuan He, Xirui Hu, Yanbo Ding, Yefei He, Yifan Yang, Youping Gu, Yuqing Yang, Zeyu Zhang, Zhiyuan He

Reinforcement learning with feedback from 3D models enforces geometric consistency in text-to-video generation without changing the base architecture.

arxiv:2604.24764 v2 · 2026-04-27 · cs.CV

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{KKLSRPH4474HO5EJ6H5QDCT6FE}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

Utilizing Flow-GRPO, we optimize the model using feedback from pre-trained 3D foundation models and vision-language models to enforce structural coherence without altering the underlying architecture... Extensive evaluations reveal that our approach significantly enhances 3D consistency while preserving the original visual quality of the foundation model.

C2weakest assumption

That feedback signals from pre-trained 3D foundation models and vision-language models provide reliable, unbiased measures of structural coherence that translate directly to improved video generation.

C3one line summary

World-R1 uses RL with 3D model feedback and a new text dataset to improve geometric consistency in text-to-video generation while keeping the base model unchanged.

Cited by

2 papers in Pith

LaMo: Self-Supervised Latent Motion Priors for Physical Realism in Video Generation

Geo-Align: Video Generation Alignment via Metric Geometry Reward

Receipt and verification

First computed	2026-05-21T01:05:19.520643Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

529728bcfce7f8777489f1fb018a7e29199d2ccacdf51b57c12bef710d506bd1

Aliases

arxiv: 2604.24764 · arxiv_version: 2604.24764v2 · doi: 10.48550/arxiv.2604.24764 · pith_short_12: KKLSRPH4474H · pith_short_16: KKLSRPH4474HO5EJ · pith_short_8: KKLSRPH4

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/KKLSRPH4474HO5EJ6H5QDCT6FE \
  | 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: 529728bcfce7f8777489f1fb018a7e29199d2ccacdf51b57c12bef710d506bd1

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "00db5d85e5165711255697223170b23f4b7cf47aeef3efc6ea48f8aed1397dd8",
    "cross_cats_sorted": [],
    "license": "http://creativecommons.org/licenses/by-nc-sa/4.0/",
    "primary_cat": "cs.CV",
    "submitted_at": "2026-04-27T17:59:56Z",
    "title_canon_sha256": "2b180adec08e74911f06561a9ed78579237c978d0e1d663a766bb5fb5141dba3"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2604.24764",
    "kind": "arxiv",
    "version": 2
  }
}