Pith Number

pith:ABOPGO4I

pith:2026:ABOPGO4IOWB3A3OTJQASI5TPYD

not attested not anchored not stored refs resolved

GraphThinker: Reinforcing Temporally Grounded Video Reasoning with Event Graph Thinking

Da Li, Jian Hu, Shaogang Gong, Wei Li, Yuhang Zang, Ziquan Liu, Zixu Cheng

GraphThinker builds event graphs and applies visual rewards in reinforcement finetuning to ground MLLM video reasoning and reduce temporal hallucinations.

arxiv:2602.17555 v3 · 2026-02-19 · cs.CV

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

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

GraphThinker achieves an over 4% improvement in IoU=0.3 for moment localisation on RexTime and a 9.8% improvement in reducing temporal sequence hallucination plus 7.6% gain in Binary QA for action hallucination on VidHalluc compared to state-of-the-art methods.

C2weakest assumption

That an MLLM can reliably construct an accurate Event-based Video Scene Graph capturing true intra- and inter-event relations without introducing its own hallucinations or errors.

C3one line summary

GraphThinker reduces temporal hallucinations in video reasoning by constructing event-based scene graphs and applying visual attention rewards in reinforcement finetuning.

References

74 extracted · 74 resolved · 13 Pith anchors

[1] The claude 3 model family: Opus, sonnet, haiku.Claude-3 Model Card, 1(1):4, 2024 2024

[2] Qwen2.5-VL Technical Report 2025 · arXiv:2502.13923

[3] arXiv preprint arXiv:2503.06486 (2025)

[4] Rextime: A benchmark suite for reasoning-across-time in videos.Advances in Neural In- formation Processing Systems, 37:28662–28673, 2024 2024

[5] Sharegpt4video: Improving video understand- ing and generation with better captions.Advances in Neural Information Processing Systems, 37:19472–19495, 2024 2024

Receipt and verification

First computed	2026-05-18T02:44:31.104022Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

005cf33b887583b06dd34c0124766fc0d734630627c5bf474c4457dfd2ddf59e

Aliases

arxiv: 2602.17555 · arxiv_version: 2602.17555v3 · doi: 10.48550/arxiv.2602.17555 · pith_short_12: ABOPGO4IOWB3 · pith_short_16: ABOPGO4IOWB3A3OT · pith_short_8: ABOPGO4I

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/ABOPGO4IOWB3A3OTJQASI5TPYD \
  | 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: 005cf33b887583b06dd34c0124766fc0d734630627c5bf474c4457dfd2ddf59e

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "738bbbadb35f4aaa474394bc167435c3e2efc39fbc041992bcf5b04ccb0614c0",
    "cross_cats_sorted": [],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.CV",
    "submitted_at": "2026-02-19T17:09:30Z",
    "title_canon_sha256": "1d065cabbd4798020bb1a5f91a4a1cc24944fc30f3020b58d068a8196b4cd80d"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2602.17555",
    "kind": "arxiv",
    "version": 3
  }
}