测试视觉语言模型理解人类意图的能力，发现小模型也能准确推理但易被误导。
How Well Can Vison-Language Models Understand Humans' Intention? An Open-ended Theory of Mind Question Evaluation Benchmark
- 构建开放问答框架，评估模型在多种心理状态推理任务中的表现
- GPT-4 最优，仅 GPT-4o-mini 表现接近，小模型常误用视觉线索
- 聚焦欺凌、作弊等复杂场景，揭示模型在真实意图理解上的局限
视觉语言模型（VLMs）在视觉问答任务中展现出强大的推理能力，但在理论心智（ToM）任务——如推断人类意图、信念和心理状态——方面仍缺乏深入探索。本文提出一种开放式问题框架，用于评估VLM在多样化ToM任务中的表现。我们收集并标注了包含30张图像的基准数据集，并评估了四种不同规模的VLM。结果显示，GPT-4表现最佳，仅较小的GPT-4o-mini模型达到相近水平。研究发现，当涉及欺凌或欺骗等复杂情境时，模型普遍难以正确推断意图。有趣的是，部分小模型虽依赖错误的视觉线索，却仍能得出正确结论。数据集已开源：https://github.com/ximingwen/ToM-AAAI25-Multimodal。
原文摘要 · Abstract (English)
Vision Language Models (VLMs) have demonstrated strong reasoning capabilities in Visual Question Answering (VQA) tasks; however, their ability to perform Theory of Mind (ToM) tasks, such as inferring human intentions, beliefs, and mental states, remains underexplored. We propose an open-ended question framework to evaluate VLMs' performance across diverse categories of ToM tasks. We curated and annotated a benchmark dataset of 30 images and evaluated the performance of four VLMs of varying sizes. Our results show that the GPT-4 model outperformed all the others, with only one smaller model, GPT-4o-mini, achieving comparable performance. We observed that VLMs often struggle to infer intentions in complex scenarios such as bullying or cheating. Our findings reveal that smaller models can sometimes infer correct intentions despite relying on incorrect visual cues. The dataset is available at https://github.com/ximingwen/ToM-AAAI25-Multimodal.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。