Pith Number

pith:HBQNWEDY

pith:2026:HBQNWEDYQHMKTIPES7MNUPYUEU

not attested not anchored not stored refs resolved

PDCR: Perception-Decomposed Confidence Reward for Vision-Language Reasoning

Chang D. Yoo, Chong Luo, Eunseop Yoon, Gwanhyeong Koo, Hee Suk Yoon, Ji Woo Hong, Mark Hasegawa-Johnson, Qi Dai, SooHwan Eom

Decomposing confidence rewards into perception and reasoning clusters improves vision-language model training.

arxiv:2605.13467 v1 · 2026-05-13 · cs.CL

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{HBQNWEDYQHMKTIPES7MNUPYUEU}

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

We demonstrate that PDCR outperforms the naive, global-reward formulation and sparse-reward baselines on key V-L reasoning benchmarks.

C2weakest assumption

The unsupervised clustering driven by the model-internal Visual Dependence Score accurately partitions steps into perception and reasoning without any labeled supervision or external verification.

C3one line summary

PDCR improves vision-language reasoning by computing separate normalized confidence advantages for perception steps and reasoning steps after unsupervised decomposition.

References

79 extracted · 79 resolved · 12 Pith anchors

[1] Gpt-4v(ision) system card. 2023. 1 2023

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

[3] Vrprm: Process reward modeling via visual reasoning.arXiv preprint arXiv:2508.03556, 2025 2025

[4] arXiv preprint arXiv:2505.23558 , year= 2025

[5] Ultrafeedback: Boosting language models with scaled ai feedback

Receipt and verification

First computed	2026-05-18T02:44:41.611045Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

3860db107881d8a9a1e497d8da3f142520076fb5acc35d29893e75cd61db867b

Aliases

arxiv: 2605.13467 · arxiv_version: 2605.13467v1 · doi: 10.48550/arxiv.2605.13467 · pith_short_12: HBQNWEDYQHMK · pith_short_16: HBQNWEDYQHMKTIPE · pith_short_8: HBQNWEDY

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/HBQNWEDYQHMKTIPES7MNUPYUEU \
  | 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: 3860db107881d8a9a1e497d8da3f142520076fb5acc35d29893e75cd61db867b

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "65fbf67244486fa2246b4d2fada7fcf3f61708833de83bfe592dc929f3070947",
    "cross_cats_sorted": [],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.CL",
    "submitted_at": "2026-05-13T12:55:18Z",
    "title_canon_sha256": "3e63b5f0621ec24924545486c7cccd4d1e39f027abe3486fe1716415d604e51d"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.13467",
    "kind": "arxiv",
    "version": 1
  }
}