Date
AI Overview
Why it matters
Multimodal models struggle to evaluate artistic expression because aesthetic concepts are abstract and open-ended, and multimodal artwork annotations are scarce.
Study at a glance
- Question: Can multimodal models be trained to evaluate children's artwork across multiple dimensions and provide formative feedback?
- Setting and data: Over 1,000 artworks by children ages 5 to 15, annotated by 12 expert educators.
- Method: Benchmark creation and model evaluation
- AI tools: Qwen2.5-VL-7B
Key findings
- Applying an attribute-specific training approach increased the model's correlation with expert scores from 0.468 to 0.653 on Qwen2.5-VL-7B.
- The fine-tuning method produced the largest performance gains on perceptual evaluation dimensions.
- Attribute-aware training narrowed performance gaps on higher-order artistic evaluation attributes.
Summary drafted with AI from the paper’s abstract and may contain errors. See the full paper for details.
Multimodal Large Language Models (MLLMs) show remarkable progress across many visual-language tasks; however, their capacity to evaluate artistic expression remains limited. Aesthetic concepts are inherently abstract and open-ended, and multimodal artwork annotations are scarce. We introduce KidsArtBench, a new benchmark of over 1k children's artworks (ages 5-15) annotated by 12 expert educators across 9 rubric-aligned dimensions, together with expert comments for feedback. Unlike prior aesthetic datasets that provide single scalar scores on adult imagery, KidsArtBench targets children's artwork and pairs multi-dimensional annotations with comment supervision to enable both ordinal assessment and formative feedback. Building on this resource, we propose an attribute-specific multi-LoRA approach, where each attribute corresponds to a distinct evaluation dimension (e.g., Realism, Imagination) in the scoring rubric, with Regression-Aware Fine-Tuning (RAFT) to align predictions with ordinal scales. On Qwen2.5-VL-7B, our method increases correlation from 0.468 to 0.653, with the largest gains on perceptual dimensions and narrowed gaps on higher-order attributes. These results show that educator-aligned supervision and attribute-aware training yield pedagogically meaningful evaluations and establish a rigorous testbed for sustained progress in educational AI. We release data and code with ethics documentation.
Research Categories
What is the application?
Who is the user?
Who age?
Why use AI?
Study design
AI tools
Datasets
Publication Source
Publisher
arXiv
