arXivDaily arXiv每日学术速递 周一至周五更新
​通知:联邦节假日调休:哥伦布日 / 原住民日(Columbus Day / Indigenous Peoples' Day),今日(13日)无更新。
arXiv 2602.07532cs.CVcs.LG

评估超越物体发现的对象中心模型

Evaluating Object-Centric Models beyond Object Discovery

Department of Computer Science, TU Darmstadt(图宾根大学计算机科学系);Zuse School ELIZA(Zuse ELIZA学校)

Krishnakant Singh, Simone Schaub-Meyer, Stefan Roth

AI总结:

本文提出统一评估方法,通过指令微调的VLMs评估OCL模型在复杂推理任务中的表现,并引入统一指标解决定位与表示有用性评估的不一致问题。

摘要:

对象中心学习(OCL)旨在学习支持组合泛化和对分布外(OOD)数据鲁棒性的结构化场景表示。然而,OCL模型往往没有针对这些目标进行评估。相反,大多数先前工作仅通过物体发现和简单的推理任务来评估OCL模型,例如通过图像分类探测表示。我们识别出现有基准的两个局限性:(1)它们对OCL模型表示有用性的洞察有限;(2)定位和表示有用性使用不同的指标进行评估。为解决(1),我们使用指令微调的VLMs作为评估者,使能够在多样化的VQA数据集上进行可扩展的基准测试,以测量VLMs如何利用OCL表示进行复杂推理任务。为解决(2),我们引入了一个统一的评估任务和指标,同时评估定位(在哪里)和表示有用性(是什么),从而消除由不同评估引入的不一致。最后,我们包含一个简单的多特征重建基线作为参考点。

英文摘要:

Object-centric learning (OCL) aims to learn structured scene representations that support compositional generalization and robustness to out-of-distribution (OOD) data. However, OCL models are often not evaluated regarding these goals. Instead, most prior work focuses on evaluating OCL models solely through object discovery and simple reasoning tasks, such as probing the representation via image classification. We identify two limitations in existing benchmarks: (1) They provide limited insights on the representation usefulness of OCL models, and (2) localization and representation usefulness are assessed using disjoint metrics. To address (1), we use instruction-tuned VLMs as evaluators, enabling scalable benchmarking across diverse VQA datasets to measure how well VLMs leverage OCL representations for complex reasoning tasks. To address (2), we introduce a unified evaluation task and metric that jointly assess localization (where) and representation usefulness (what), thereby eliminating inconsistencies introduced by disjoint evaluation. Finally, we include a simple multi-feature reconstruction baseline as a reference point.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑