LinSlot: Exploiting Linear Representation hypothesis for unsupervised attribute discovery from slot based object representation

Computer Science > Computer Vision and Pattern Recognition arXiv:2610.10722 (cs) [Submitted on 7 Oct 2026] Title: LinSlot: Exploiting Linear Representation hypothesis for unsupervised attribute discovery from slot based object representation Authors: Sanket Gandhi, Utkarsh Giri, Varun Subramanium, Rohan Paul, Parag Singla.

计算机科学 > 计算机视觉与模式识别 arXiv:2610.10722 (cs) [2026年10月7日提交] 标题：LinSlot：利用线性表示假设从基于槽的对象表示中进行无监督属性发现 作者：Sanket Gandhi, Utkarsh Giri, Varun Subramanium, Rohan Paul, Parag Singla。

Abstract: This paper studies the problem of learning disentangled representations of objects and their attributes from raw, unstructured image data. Slot-based methods have shown considerable success in unsupervised learning of object representations from images. Block-slot attention-based methods extend this framework to attribute representations by assuming a uniform factorization of object representations into attributes, which may be suboptimal and consequently limit the quality of the learned representations.

摘要：本文研究了从原始、非结构化的图像数据中学习对象及其属性的解耦表示的问题。基于槽（Slot-based）的方法在从图像中无监督学习对象表示方面已显示出相当大的成功。基于块-槽（Block-slot）注意力的方法通过假设对象表示到属性的统一分解，将该框架扩展到属性表示，但这可能是次优的，并因此限制了所学习表示的质量。

We therefore investigate a framework for jointly discovering object and attribute representations. Our key contribution is leveraging the Linear Representation Hypothesis (LRH), which postulates that composable concepts can be represented as linearly additive subspaces in slot representations. Based on this insight, we propose a probabilistic model connecting images, slots (objects), and blocks (attributes).

因此，我们研究了一个联合发现对象和属性表示的框架。我们的主要贡献是利用了线性表示假设（LRH），该假设认为可组合的概念可以在槽表示中表示为线性加性子空间。基于这一见解，我们提出了一个连接图像、槽（对象）和块（属性）的概率模型。

We present an architecture that leverages block attention to connect attribute representations to slots and incorporates LRH in both object and attribute representation spaces. This architecture effectively optimizes the Evidence Lower Bound (ELBO) of the proposed graphical model. Our experiments demonstrate (i) effective discovery of disentangled object and attribute representations, (ii) empirical evidence for LRH in slot space, and (iii) the ability to perform image editing owing to the disentangled and interpretable nature of the learned representations.

我们提出了一种架构，利用块注意力将属性表示连接到槽，并在对象和属性表示空间中结合了LRH。该架构有效地优化了所提图形模型的证据下界（ELBO）。我们的实验证明了：（i）有效发现了对象和属性的解耦表示；（ii）在槽空间中为LRH提供了实证证据；（iii）由于所学表示的解耦和可解释性，能够进行图像编辑。

Our experiments on multiple datasets demonstrate improvements in DCI scores over state-of-the-art methods. Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI).

我们在多个数据集上的实验表明，DCI分数较现有最先进方法有所提高。学科：计算机视觉与模式识别 (cs.CV)；人工智能 (cs.AI)。