Identifiable Object Representations under Spatial Ambiguities

TL;DR

This paper presents a multi-view probabilistic method for object representations that effectively resolves spatial ambiguities like occlusions and view ambiguities, achieving invariance and disentanglement without viewpoint labels.

Contribution

It introduces a novel multi-view probabilistic framework that guarantees identifiability and disentanglement of object content and viewpoint information without requiring viewpoint annotations.

Findings

Resolves spatial ambiguities in object representations.

Provides theoretical guarantees for identifiability.

Demonstrates robustness and scalability on benchmarks.

Abstract

Modular object-centric representations are essential for *human-like reasoning* but are challenging to obtain under spatial ambiguities, *e.g. due to occlusions and view ambiguities*. However, addressing challenges presents both theoretical and practical difficulties. We introduce a novel multi-view probabilistic approach that aggregates view-specific slots to capture *invariant content* information while simultaneously learning disentangled global *viewpoint-level* information. Unlike prior single-view methods, our approach resolves spatial ambiguities, provides theoretical guarantees for identifiability, and requires *no viewpoint annotations*. Extensive experiments on standard benchmarks and novel complex datasets validate our method's robustness and scalability.