CodePercept: Code-Grounded Visual STEM Perception for MLLMs | Signal Canvas | ScienceToStartup

← Back to Paper

CodePercept: Code-Grounded Visual STEM Perception for MLLMs

Stale68d agoVerification pending / evidence receipt incomplete

Export Brief Open in Build Loop Connect with Author

Viability

0.0/10

Compared to this week’s papers

Verification pending

Use This Via API or MCP

Use Signal Canvas as the narrative proof surface

Signal Canvas is the citation-first public layer for turning one paper into a structured commercialization narrative. Use it to hand off into REST, MCP, Build Loop, and launch-pack execution without losing source lineage.

Signal Canvas API Paper Proof Page Open Build Loop Launch Pack Example

Use This Via API or MCP

Use this Signal Canvas via API or MCP

Route this paper proof surface into REST, MCP, or developer workflows while preserving the same evidence receipt and related-resource context.

Signal Canvas guide REST guide MCP guide

Page Freshness

Signal Canvas proof surface

Canonical route: /signal-canvas/codepercept-code-grounded-visual-stem-perception-for-mllms

stale

Proof freshness: stale
Proof status: unverified
Display score: 8/10
Last proof check: 2026-04-02
Score updated: 2026-04-02
Score fresh until: 2026-05-02
References: 0
Source count: 0
Coverage: 17%

This page is showing the last landed evidence receipt and score bundle because the latest proof data is outside the freshness window.

Agent Handoff

CodePercept: Code-Grounded Visual STEM Perception for MLLMs

Canonical ID codepercept-code-grounded-visual-stem-perception-for-mllms | Route /signal-canvas/codepercept-code-grounded-visual-stem-perception-for-mllms

REST example

curl https://sciencetostartup.com/api/v1/agent-handoff/signal-canvas/codepercept-code-grounded-visual-stem-perception-for-mllms

MCP example

{
  "tool": "search_signal_canvas",
  "arguments": {
    "mode": "paper",
    "paper_ref": "codepercept-code-grounded-visual-stem-perception-for-mllms",
    "query_text": "Summarize CodePercept: Code-Grounded Visual STEM Perception for MLLMs"
  }
}

source_context

{
  "surface": "signal_canvas",
  "mode": "paper",
  "query": "CodePercept: Code-Grounded Visual STEM Perception for MLLMs",
  "normalized_query": "2603.10757",
  "route": "/signal-canvas/codepercept-code-grounded-visual-stem-perception-for-mllms",
  "paper_ref": "codepercept-code-grounded-visual-stem-perception-for-mllms",
  "topic_slug": null,
  "benchmark_ref": null,
  "dataset_ref": null
}

Paper mode· single-doc scopescope: codepercept-code-grounded-visual-stem-perception-for-mllms

Preparing verified analysis

GitHub Code Pulse

No public code linked for this paper yet.

Claim map

Strong 8Mixed 0Weak 0

Evidencepartial
scaling perception consistently outperforms scaling reasoning
Implicationpartial
Directly stated in abstract as a critical insight from systematic scaling analysis
Verificationpartial
partial
Evidencepartial
reveals perception as the true lever limiting current STEM visual reasoning
Implicationpartial
Explicitly stated conclusion from systematic scaling analysis
Verificationpartial
partial
Evidencepartial
executable code provides precise semantics that naturally align with the structured nature of STEM visuals
Implicationpartial
Directly stated rationale for using code as perceptual medium
Verificationpartial
partial
Evidencepartial
we construct ICC-1M, a large-scale dataset comprising 1M Image-Caption-Code triplets
Implicationpartial
Explicit numeric claim about dataset size and composition
Verificationpartial
partial
Evidencepartial
Code-Grounded Caption Generation treats executable code as ground truth for image captions, eliminating the hallucinations inherent in existing knowledge distillation methods
Implicationpartial
Direct claim about method advantage with specific technical justification
Verificationpartial
partial
Evidencepartial
STEM Image-to-Code Translation prompts models to generate reconstruction code, mitigating the ambiguity of natural language for perception enhancement
Implicationpartial
Direct claim about method mechanism and benefit
Verificationpartial
partial
Evidencepartial
our benchmark requires comprehensive visual comprehension through executable code generation for image reconstruction, providing deterministic and verifiable assessment
Implicationpartial
Direct claim about benchmark characteristics and advantages
Verificationpartial
partial
Evidencepartial
Unlike existing work relying on problem-solving accuracy as a proxy that only measures problem-relevant understanding
Implicationpartial
Implied limitation of existing approaches stated in comparison to new benchmark
Verificationpartial
partial

Startup potential card

Startup potential card preview

Share on X LinkedIn