CURV: Enhancing Chart Understanding Through Curriculum Visual Grounded Reasoning
Xuehang Guo, Pingyue Zhang, Ruiyi Zhang, Zhenhailong Wang, Hanrui Lyu, Heng Ji, Tong Sun, Qingyun Wang, Manling Li
cs.CV, cs.AI, cs.CL
Submitted: 2026-08-03
Project page: https://xhguo7.github.io/CURV
License: http://creativecommons.org/licenses/by/4.0/
The gist: Chart question answering (CQA) requires multimodal large language models (MLLMs) to integrate visual comprehension with logical reasoning, yet current models struggle with accurate visual grounding
Terminology
Abstract
Chart question answering (CQA) requires multimodal large language models (MLLMs) to integrate visual comprehension with logical reasoning, yet current models struggle with accurate visual grounding and coherent reasoning chains. While extrinsic chain-of-thought prompting and visual cues significantly improve performance, current MLLMs lack intrinsic visual grounded reasoning capabilities, leading to inaccurate perception and reasoning disconnected from visual evidence. To address these limitations, we propose CURV, a curriculum learning framework that develops intrinsic visual reasoning capabilities by reformulating CQA as multi-step visual grounded reasoning, where each step coordinates logical reasoning with dynamic visual grounding through spatial attention concentration. To assist model learning, we further introduce CCQA, a three-level curriculum dataset with scalable synthetic generation across diverse chart types and reasoning patterns. Our curriculum systematically progresses from basic single-operation reasoning to complex multi-chart compositional tasks. Experiments demonstrate that CURV achieves up to 20.50% improvements over baselines and is generalizable to real-world benchmarks (up to 12.30%) and out-of-domain multimodal reasoning tasks (up to 10.20%), validating the effectiveness of internalizing visual reasoning with dynamic grounding for enhanced chart understanding capabilities. Code is available at: https://xhguo7.github.io/CURV/.
Sources
- SAM 3: Segment Anything with Concepts
- MINT-CoT: Enabling Interleaved Visual Tokens in Mathematical Chain-of-Thought Reasoning
- Limits to scalable evaluation at the frontier: LLM as Judge won't beat twice the data
- Navigating the Digital World as Humans Do: Universal Visual Grounding for GUI Agents
- SIMPLOT: Enhancing Chart Question Answering by Distilling Essentials
- CalibraEval: Calibrating Prediction Distribution to Mitigate Selection Bias in LLMs-as-Judges
- Synthesize Step-by-Step: Tools, Templates and LLMs as Data Generators for Reasoning-Based Chart VQA
- MathVista: Evaluating Mathematical Reasoning of Foundation Models in Visual Contexts
- ChartGemma: Visual Instruction-tuning for Chart Reasoning in the Wild
- CogCoM: A Visual Language Model with Chain-of-Manipulations Reasoning
- Qwen2.5 Technical Report
- Visual CoT: Advancing Multi-Modal Language Models with a Comprehensive Dataset and Benchmark for Chain-of-Thought Reasoning
- ChartMuseum: Testing Visual Reasoning Capabilities of Large Vision-Language Models
- Traceable Evidence Enhanced Visual Grounded Reasoning: Evaluation and Methodology
- VGR: Visual Grounded Reasoning
- Perception-Aware Policy Optimization for Multimodal Reasoning
- Visually Descriptive Language Model for Vector Graphics Reasoning
- CharXiv: Charting Gaps in Realistic Chart Understanding in Multimodal LLMs
- Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
- GUI-Actor: Coordinate-Free Visual Grounding for GUI Agents
Related papers
- Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
- AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
- Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
- MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
- TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
- A Survey on Efficient Vision-Language-Action Models