Language-Mediated, Object-Centric Representation Learning

TL;DR

LORL introduces a novel framework that combines vision and language to learn object-centric scene representations, improving unsupervised object discovery and aiding downstream tasks like referring expression comprehension.

Contribution

LORL is the first to integrate language concepts with unsupervised object discovery algorithms, enhancing their performance and interpretability.

Findings

LORL improves object discovery accuracy on benchmark datasets.

Language integration enhances concept learning and downstream task performance.

LORL is compatible with various unsupervised object discovery methods.

Abstract

We present Language-mediated, Object-centric Representation Learning (LORL), a paradigm for learning disentangled, object-centric scene representations from vision and language. LORL builds upon recent advances in unsupervised object discovery and segmentation, notably MONet and Slot Attention. While these algorithms learn an object-centric representation just by reconstructing the input image, LORL enables them to further learn to associate the learned representations to concepts, i.e., words for object categories, properties, and spatial relationships, from language input. These object-centric concepts derived from language facilitate the learning of object-centric representations. LORL can be integrated with various unsupervised object discovery algorithms that are language-agnostic. Experiments show that the integration of LORL consistently improves the performance of unsupervised object discovery methods on two datasets via the help of language. We also show that concepts learned by LORL, in conjunction with object discovery methods, aid downstream tasks such as referring expression comprehension.