The first benchmark for AI-generated scientific figure detection shows existing detectors fail in zero-shot transfer, overfit to specific generators, and break under common image corruptions.
You are an expert on translating academic writing to visual specification
10 Pith papers cite this work, alongside 1 external citations. Polarity classification is still indexing.
10
Pith papers citing it
1
external citations · external index
citation-role summary
background 1
citation-polarity summary
years
2026 10roles
background 1polarities
background 1representative citing papers
LLM-native figures embed provenance and enable direct LLM interaction with scientific visualizations to accelerate discovery and improve reproducibility.
GENFIG1 is a new benchmark that tests whether vision-language models can create effective Figure 1 visuals capturing the central scientific idea from paper text.
DiagramBank is a large-scale curated dataset of 89,422 schematic diagrams from scientific papers with rich metadata to support multimodal retrieval and exemplar-driven figure generation.
SAGE with MHFA improves failure recovery in autonomous research agents, raising metrics-bearing outputs from 42% to 92% on a 12-topic benchmark versus single-reflection baselines.
Introduces SciIR-82k dataset and SciIR-Bench for scientific image reasoning generation organized by Peirce's semiotic triad, with fine-tuning raising model score from 35% to 43%.
Introduces FEPBench benchmark to evaluate T2I models on instruction faithfulness, reasoning enrichment, and semantic precision for natural-science illustrations using atom set annotations.
DiagramRAG is a retrieval-augmented framework that represents diagrams as knowledge graphs, synthesizes sketch variants, trains an embedding model for structure-aware retrieval, and uses retrieved references to guide sketch-based scientific diagram generation.
AI can generate research artifacts faster than it can verify them, so across all eight lifecycle stages the credible deployment mode is human-governed collaboration rather than full autonomy.
citing papers explorer
-
SciFigDetect: A Benchmark for AI-Generated Scientific Figure Detection
The first benchmark for AI-generated scientific figure detection shows existing detectors fail in zero-shot transfer, overfit to specific generators, and break under common image corruptions.
-
Figures as Interfaces: Toward LLM-Native Artifacts for Scientific Discovery
LLM-native figures embed provenance and enable direct LLM interaction with scientific visualizations to accelerate discovery and improve reproducibility.
-
GENFIG1: Visual Summaries of Scholarly Work as a Challenge for Vision-Language Models
GENFIG1 is a new benchmark that tests whether vision-language models can create effective Figure 1 visuals capturing the central scientific idea from paper text.
-
DiagramBank: A Large-scale Dataset of Diagram Design Exemplars with Paper Metadata for Retrieval-Augmented Generation
DiagramBank is a large-scale curated dataset of 89,422 schematic diagrams from scientific papers with rich metadata to support multimodal retrieval and exemplar-driven figure generation.
-
One Reflection Is Not Enough: Self-Correcting Autonomous Research via Multi-Hypothesis Failure Attribution
SAGE with MHFA improves failure recovery in autonomous research agents, raising metrics-bearing outputs from 42% to 92% on a 12-topic benchmark versus single-reflection baselines.
-
SciIR: A Large-scale Training Dataset and Benchmark for Scientific Image Reasoning Generation
Introduces SciIR-82k dataset and SciIR-Bench for scientific image reasoning generation organized by Peirce's semiotic triad, with fine-tuning raising model score from 35% to 43%.
-
Faithful, Enriched, and Precise: Benchmarking Natural-Science Illustration Generation by T2I models
Introduces FEPBench benchmark to evaluate T2I models on instruction faithfulness, reasoning enrichment, and semantic precision for natural-science illustrations using atom set annotations.
-
DiagramRAG: A Lightweight Framework to Retrieve Scientific Diagram for Figure Generation
DiagramRAG is a retrieval-augmented framework that represents diagrams as knowledge graphs, synthesizes sketch variants, trains an embedding model for structure-aware retrieval, and uses retrieved references to guide sketch-based scientific diagram generation.
-
AI for Auto-Research: Roadmap & User Guide
AI can generate research artifacts faster than it can verify them, so across all eight lifecycle stages the credible deployment mode is human-governed collaboration rather than full autonomy.
- Personalization as Inverse Planning: Learning Latent Design Intents for Agentic Slide Generation via Structural Denoising