做多模态嵌入或视觉-语言模型研究的开发者，这篇论文点出了 CLIP 类模型在概念绑定上的核心瓶颈，并给出了可复现的解决方案，值得深入阅读。
人类能轻松将颜色与形状绑定（如“红色圆形”），但 CLIP 等视觉-语言嵌入模型在多物体场景中无法正确绑定概念，表现为“词袋”行为。研究发现，CLIP 的场景嵌入可加性分解为物体表征，但绑定函数复杂度高，导致图像与文本编码器无法学习共享的绑定机制，难以泛化到未见过的概念组合。通过从头训练的受控 Transformer 模型，作者发现当数据覆盖足够时，模型能学会低复杂度的绑定函数（涉及概念间的乘法交互），实现系统性泛化。该工作揭示了嵌入模型在概念绑定上的根本限制与突破条件，代码已开源。
How can embedding models bind concepts?
Humans easily determine which color belongs to which shape in multi-object scenes, an ability known as concept binding. Vision-language embedding models such as CLIP struggle with binding: they recognize individual concepts but fail to represent which concepts form which objects. Although CLIP behaves like a bag-of-concepts model in cross-modal retrieval, object information is recoverable from its image and text embeddings separately. We study this tension through the binding function, which maps concepts to scene embeddings. We find that scene embeddings decompose additively into object representations, explaining why uni-modal probes can recover object information. However, CLIP's binding function is high-complexity, which likely prevents the image and text encoders from learning a shared binding mechanism that generalizes to unseen concept combinations. We then ask whether this limitation is fundamental. We show that it is not. In controlled transformer models trained from scratch, binding generalization emerges with sufficient data coverage. These models learn low-complexity binding functions characterized by multiplicative interactions between concepts, enabling systematic generalization. Code is publicly available at https://github.com/oshapio/binding-concepts-complexity.