CaltechAUTHORS
  A Caltech Library Service

Re-ViLM: Retrieval-Augmented Visual Language Model for Zero and Few-Shot Image Captioning

Yang, Zhuolin and Ping, Wei and Liu, Zihan and Korthikanti, Vijay and Nie, Weili and Huang, De-An and Fan, Linxi and Yu, Zhiding and Lan, Shiyi and Li, Bo and Liu, Ming-Yu and Zhu, Yuke and Shoeybi, Mohammad and Catanzaro, Bryan and Xiao, Chaowei and Anandkumar, Anima (2023) Re-ViLM: Retrieval-Augmented Visual Language Model for Zero and Few-Shot Image Captioning. . (Unpublished) https://resolver.caltech.edu/CaltechAUTHORS:20230316-153738224

[img] PDF - Submitted Version
Creative Commons Attribution.

1MB

Use this Persistent URL to link to this item: https://resolver.caltech.edu/CaltechAUTHORS:20230316-153738224

Abstract

Augmenting pretrained language models (LMs) with a vision encoder (e.g., Flamingo) has obtained state-of-the-art results in image-to-text generation. However, these models store all the knowledge within their parameters, thus often requiring enormous model parameters to model the abundant visual concepts and very rich textual descriptions. Additionally, they are inefficient in incorporating new data, requiring a computational-expensive fine-tuning process. In this work, we introduce a Retrieval-augmented Visual Language Model, Re-ViLM, built upon the Flamingo, that supports retrieving the relevant knowledge from the external database for zero and in-context few-shot image-to-text generations. By storing certain knowledge explicitly in the external database, our approach reduces the number of model parameters and can easily accommodate new data during evaluation by simply updating the database. We also construct an interleaved image and text data that facilitates in-context few-shot learning capabilities. We demonstrate that Re-ViLM significantly boosts performance for image-to-text generation tasks, especially for zero-shot and few-shot generation in out-of-domain settings with 4 times less parameters compared with baseline methods.


Item Type:Report or Paper (Discussion Paper)
Related URLs:
URLURL TypeDescription
http://arxiv.org/abs/2302.04858arXivDiscussion Paper
ORCID:
AuthorORCID
Huang, De-An0000-0002-6945-7768
Fan, Linxi0000-0001-7393-3125
Yu, Zhiding0000-0003-1776-996X
Liu, Ming-Yu0000-0002-2951-2398
Zhu, Yuke0000-0002-9198-2227
Catanzaro, Bryan0000-0003-0034-7728
Xiao, Chaowei0000-0002-7043-4926
Anandkumar, Anima0000-0002-6974-6797
Additional Information:Attribution 4.0 International (CC BY 4.0)
Record Number:CaltechAUTHORS:20230316-153738224
Persistent URL:https://resolver.caltech.edu/CaltechAUTHORS:20230316-153738224
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:120086
Collection:CaltechAUTHORS
Deposited By: George Porter
Deposited On:16 Mar 2023 20:15
Last Modified:16 Mar 2023 20:15

Repository Staff Only: item control page