Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Rose, Daniel, Himakunthala, Vaishnavi, Ouyang, Andy, He, Ryan, Mei, Alex, Lu, Yujie, Saxon, Michael, Sonar, Chinmay, Mirza, Diba, Wang, William Yang
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2305.02317
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866909080047583232
author Rose, Daniel
Himakunthala, Vaishnavi
Ouyang, Andy
He, Ryan
Mei, Alex
Lu, Yujie
Saxon, Michael
Sonar, Chinmay
Mirza, Diba
Wang, William Yang
author_facet Rose, Daniel
Himakunthala, Vaishnavi
Ouyang, Andy
He, Ryan
Mei, Alex
Lu, Yujie
Saxon, Michael
Sonar, Chinmay
Mirza, Diba
Wang, William Yang
contents Recent advances in large language models elicit reasoning in a chain-of-thought that allows models to decompose problems in a human-like fashion. Though this paradigm improves multi-step reasoning ability in language models, it is limited by being unimodal and applied mainly to question-answering tasks. We claim that incorporating visual augmentation into reasoning is essential, especially for complex, imaginative tasks. Consequently, we introduce VCoT, a novel method that leverages chain-of-thought prompting with vision-language grounding to recursively bridge the logical gaps within sequential data. Our method uses visual guidance to generate synthetic multimodal infillings that add consistent and novel information to reduce the logical gaps for downstream tasks that can benefit from temporal reasoning, as well as provide interpretability into models' multi-step reasoning. We apply VCoT to the Visual Storytelling and WikiHow summarization datasets and demonstrate through human evaluation that VCoT offers novel and consistent synthetic data augmentation beating chain-of-thought baselines, which can be used to enhance downstream performance.
format Preprint
id arxiv_https___arxiv_org_abs_2305_02317
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Visual Chain of Thought: Bridging Logical Gaps with Multimodal Infillings
Rose, Daniel
Himakunthala, Vaishnavi
Ouyang, Andy
He, Ryan
Mei, Alex
Lu, Yujie
Saxon, Michael
Sonar, Chinmay
Mirza, Diba
Wang, William Yang
Computation and Language
Computer Vision and Pattern Recognition
Recent advances in large language models elicit reasoning in a chain-of-thought that allows models to decompose problems in a human-like fashion. Though this paradigm improves multi-step reasoning ability in language models, it is limited by being unimodal and applied mainly to question-answering tasks. We claim that incorporating visual augmentation into reasoning is essential, especially for complex, imaginative tasks. Consequently, we introduce VCoT, a novel method that leverages chain-of-thought prompting with vision-language grounding to recursively bridge the logical gaps within sequential data. Our method uses visual guidance to generate synthetic multimodal infillings that add consistent and novel information to reduce the logical gaps for downstream tasks that can benefit from temporal reasoning, as well as provide interpretability into models' multi-step reasoning. We apply VCoT to the Visual Storytelling and WikiHow summarization datasets and demonstrate through human evaluation that VCoT offers novel and consistent synthetic data augmentation beating chain-of-thought baselines, which can be used to enhance downstream performance.
title Visual Chain of Thought: Bridging Logical Gaps with Multimodal Infillings
topic Computation and Language
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2305.02317