AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Easy for Children, Hard for AI: The Limits of Multimodal LLMs in Early Childhood Learning

Jingping Liu, Xueyan Wu, Hanxuan Chen, Ziyan Liu, Zhangquan Chen, Ronghao Chen, Huacan Wang

AAAI 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Benchmarks and analysis of MLLMs · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Evaluates multimodal LLMs on early-childhood-style visual perception and reasoning tasks, finding them substantially weaker than the child-level skills they are meant to support.

Problem

Early childhood education is a critical developmental stage, but the ability of multimodal LLMs to support visual perception and reasoning for young learners is largely unexplored, and existing child-related AI focuses on language, emotion, or behavior.

Approach

Frames early-childhood visual perception and reasoning as an evaluation target for MLLMs, probing where current models fall short of tasks easy for children.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
LoRA fine-tune on ChildBench train (ablation)Qwen2.5-VL-7B · ChildBench testaccuracy: 32.86 -> 52.01Improved
LoRA fine-tune on ChildBench train (ablation)LLaMA-3.2-11B · ChildBench testaccuracy: 23.10 -> 25.35Improved
LoRA fine-tune on ChildBench train (ablation)LLaVA-v1.5-7B · ChildBench testaccuracy: 17.68 -> 19.08Improved
LoRA fine-tune on ChildBench train (ablation)DeepSeek-VL-7B · ChildBench testaccuracy: 17.50 -> 21.23Improved
LoRA fine-tune on ChildBench train (ablation)MiniCPM-V-8B (Visual Discrimination) · ChildBench Visual Discriminationaccuracy: 25.64 -> 52.14Improved
add one in-context example (1-shot) (ablation)GPT-4o · ChildBench (500 samples)accuracy: 37.40 -> 29.20Hurt
add one in-context example (1-shot) (ablation)Gemini-2.5-Pro · ChildBench (500 samples)accuracy: 44.60 -> 45.80Improved
scale Qwen2.5-VL from 7B to 32B (ablation)Qwen2.5-VL · ChildBench testaccuracy: 52.01 -> 55.00Improved
scale InternVL3.0 from 8B to 14B (ablation)InternVL3.0 LoRA · ChildBench testaccuracy: 41.38 -> 40.40Mixed
scale InternVL3.0 from 8B to 38B (ablation)InternVL3.0 LoRA · ChildBench testaccuracy: 41.38 -> 63.52Improved
merge multiple images into a single image (ablation)Qwen2.5-VL-7B (LoRA) · ChildBench multi-image samplesaccuracy: 43.48 -> 22.83Hurt
merge multiple images into a single image (ablation)Gemini-2.5-Pro · ChildBench multi-image samplesaccuracy: 19.40 -> 53.73Improved

Surprising findings

Limitations

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy