Enregistré dans:
Détails bibliographiques
Auteurs principaux: Luo, Jiayun, Hossain, Mir Rayat Imtiaz, Sarkar, Pritam, Li, Boyang, Sigal, Leonid
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:https://arxiv.org/abs/2412.08110
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866911655608188928
author Luo, Jiayun
Hossain, Mir Rayat Imtiaz
Sarkar, Pritam
Li, Boyang
Sigal, Leonid
author_facet Luo, Jiayun
Hossain, Mir Rayat Imtiaz
Sarkar, Pritam
Li, Boyang
Sigal, Leonid
contents Vision-Language Models (VLMs) have achieved strong performance on implicit and explicit visual grounding and related tasks. However, such abilities are generally tested on simple, single-object phrases. We find that grounding performance degrades for complex, multi-object references. These limitations largely arise from training objectives that leverage image-caption alignment, where direct multi-object references are rare, the number of possible such references is theoretically large (exponential in the number of objects), and attribution is difficult. To address this, without requiring any additional annotations, we propose Compositional Attention-Regularized Training (CompART), which decomposes captions into object-centric phrases and constructs composite phrases by pairing them with conjunctions. We then introduce a composition loss that encourages the attention induced by a composite phrase to equal the sum of the attentions of its constituent phrases, promoting balanced multi-object localization. We evaluate CompART across four VLM architectures, spanning both contrastive-based and generative-based models, on four benchmarks for multi-object grounding and two VQA benchmarks for general visual understanding. CompART consistently improves grounding for both single- and multi-object references across diverse VLM architectures and datasets, and further demonstrates enhanced visual understanding, as evidenced by gains on VQA, despite not being explicitly trained for this task.
format Preprint
id arxiv_https___arxiv_org_abs_2412_08110
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle The ART of Composition: Attention-Regularized Training for Compositional Visual Grounding
Luo, Jiayun
Hossain, Mir Rayat Imtiaz
Sarkar, Pritam
Li, Boyang
Sigal, Leonid
Computer Vision and Pattern Recognition
Computation and Language
Machine Learning
Vision-Language Models (VLMs) have achieved strong performance on implicit and explicit visual grounding and related tasks. However, such abilities are generally tested on simple, single-object phrases. We find that grounding performance degrades for complex, multi-object references. These limitations largely arise from training objectives that leverage image-caption alignment, where direct multi-object references are rare, the number of possible such references is theoretically large (exponential in the number of objects), and attribution is difficult. To address this, without requiring any additional annotations, we propose Compositional Attention-Regularized Training (CompART), which decomposes captions into object-centric phrases and constructs composite phrases by pairing them with conjunctions. We then introduce a composition loss that encourages the attention induced by a composite phrase to equal the sum of the attentions of its constituent phrases, promoting balanced multi-object localization. We evaluate CompART across four VLM architectures, spanning both contrastive-based and generative-based models, on four benchmarks for multi-object grounding and two VQA benchmarks for general visual understanding. CompART consistently improves grounding for both single- and multi-object references across diverse VLM architectures and datasets, and further demonstrates enhanced visual understanding, as evidenced by gains on VQA, despite not being explicitly trained for this task.
title The ART of Composition: Attention-Regularized Training for Compositional Visual Grounding
topic Computer Vision and Pattern Recognition
Computation and Language
Machine Learning
url https://arxiv.org/abs/2412.08110