Diagram Understanding

Follow this topic Ask about this topic

Latest papers

8 papers1 paper in the last four weeks, against 1 the four weeks before. 0.0% of all new papers.

Seeing vs. Counting: Complexity Twins Show Where Multimodal LLMs Fail on Diagrams
Nov 6, 2026Yin LiuDiagram Understanding MLLM Evaluation
Google
SciFigBench: A Benchmark for Scientific Method Figure Understanding
Oct 19, 2026Lijia Wu, Zheng Li, Yuxuan Liu +3Diagram Understanding Scientific Figure Understanding
VCIP, CS & AAIS, Nankai University · NKIARI, Shenzhen Futian · Li Auto.
Back2Struct: Making Structured Images Editable Again
Sep 29, 2026Pengyu Yan, Yixin Wu, Yunjie Tian +1Diagram Understanding Image-to-Code Generation
University at Buffalo, SUNY
Reactivating Test-Time Scaling for Plane Geometry Problem Solving
Aug 31, 2026Xiaoqiang Kang, Shengen Wu, Maizhen Ning +5Spatial Reasoning Benchmarks Diagram Understanding
School of Advanced Technology, Xi’an Jiaotong-Liverpool University · The Hong Kong University of Science and Technology (Guangzhou) · Hithink Research +2
Generative Artificial Intelligence (GenAI) to convert images of queuing networks into verifiable simulation models: an open-weight LLM workflow approach
Jul 27, 2026Thomas Monks, Alison Harper, Amy Heather +1Diagram Understanding
Text to Automata Diagrams: Comparing TikZ Code Generation with Direct Image Synthesis
Mar 9, 2026Ethan Young, Zichun Wang, Aiden Taylor +6Diagram Understanding Code Generation
Marshall University · West Virginia State University
Pseudo Contrastive Learning for Diagram Comprehension in Multimodal Models
Feb 27, 2026Hiroshi SasakiDiagram Understanding Contrastive Learning
The Japan Research Institute, Limited
Enhancing Geometric Perception in VLMs via Translator-Guided Reinforcement Learning
Feb 26, 2026Hao Yu, Shuning Jia, Guanghao Li +2Diagram Understanding RL for VLMs
Tsinghua University · Guangdong Laboratory of Artificial Intelligence and Digital Economy (SZ)

Nov 6, 2026J/K move · Enter open · S save

Seeing vs. Counting: Complexity Twins Show Where Multimodal LLMs Fail on Diagrams

Yin Liu

Google

Diagram Understanding MLLM Evaluation Counting in VLMs Multimodal Reasoning

Open paper AskPDF arXiv

When a multimodal LLM miscounts in a diagram, does it fail at seeing or at counting? Existing visual graph and puzzle benchmarks vary visual density and algorithmic hardness together, confounding the two. We pair counting questions about the same drawing that sit in different counting-complexity classes, such as its spanning trees (a determinant) and its simple paths (#P-complete), and use a transcription probe and a text-only control to separate reading from computing. We test Gemini, Gemma and Qwen3-VL models; for the Gemini models, the answer depends on how the model computes. In chain of thought, most of them read small drawings correctly and then fail at counting, at a rate set far more by the size of the answer than by its complexity class (for the primary model, the odds of an exact answer fall 2.38-fold per decade of the answer). Given code and an exact transcription, the primary model counts correctly even on 10x10 grids, and reading becomes the bottleneck: a single view misses horizontal edges, and a second reading of a rotated copy with crop checks reads 86/96 large grids exactly, against 57/96 when the upright reading is simply repeated.