Pith Number

pith:2PLSOTKM

pith:2026:2PLSOTKMJC5BHO7J2HNOK3U4QI

not attested not anchored not stored refs resolved

Bridging the Missing-Modality Gap: Improving Text-Only Calibration of Vision Language Models

Chaeyun Jang, Juho Lee (Kim Jaechul Graduate School of AI, Jungwon Choi, KAIST), Mingyeong Kim

A lightweight cross-attention module can predict missing visual embeddings from text to restore accuracy and calibration in vision-language models.

arxiv:2605.12517 v1 · 2026-04-03 · cs.CL · cs.AI · 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{2PLSOTKMJC5BHO7J2HNOK3U4QI}

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 propose the Latent Imagination Module (LIM), a lightweight cross-attention module that predicts imagined latent embeddings from textual input and feeds them into a frozen VLM backbone without pixel-level image synthesis. Across text-only benchmarks, unseen tasks, and missing-image scenarios, LIM improves accuracy and reduces calibration error.

C2weakest assumption

That cross-attention predictions of latent visual embeddings from text will be sufficiently accurate and compatible to substitute for real visual input inside a frozen VLM without introducing new systematic errors.

C3one line summary

A new Latent Imagination Module uses cross-attention to predict latent visual embeddings from text, improving accuracy and calibration of vision-language models on text-only inputs.

References

13 extracted · 13 resolved · 6 Pith anchors

[1] Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge · arXiv:1803.05457

[2] Mm-align: Learning optimal transport- based alignment dynamics for fast and accurate inference on missing modality sequences

[3] A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networks · arXiv:1610.02136

[4] Measuring Massive Multitask Language Understanding 2009 · arXiv:2009.03300

[5] GPT-4o System Card · arXiv:2410.21276

Receipt and verification

First computed	2026-05-18T03:10:02.899600Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

d3d7274d4c48ba13bbe9d1dae56e9c82223af0ca0765f4b2621990fc5d4383e3

Aliases

arxiv: 2605.12517 · arxiv_version: 2605.12517v1 · doi: 10.48550/arxiv.2605.12517 · pith_short_12: 2PLSOTKMJC5B · pith_short_16: 2PLSOTKMJC5BHO7J · pith_short_8: 2PLSOTKM

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/2PLSOTKMJC5BHO7J2HNOK3U4QI \
  | 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: d3d7274d4c48ba13bbe9d1dae56e9c82223af0ca0765f4b2621990fc5d4383e3

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "168ff8447d6fe9e085404367c568610dd7762a51ab443173cd64288d00412d73",
    "cross_cats_sorted": [
      "cs.AI",
      "cs.CV"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.CL",
    "submitted_at": "2026-04-03T10:03:02Z",
    "title_canon_sha256": "52448e062150832b7734239c8c9f7801656057475077fe50bdfba514d65e5be9"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.12517",
    "kind": "arxiv",
    "version": 1
  }
}