Skip to main content
arXiv is now an independent nonprofit! Learn more
archive
Search Submit Donate Log in
Press Enter to search · Advanced search

Computer Science > Computation and Language

arXiv:2310.05861 (cs)
[Submitted on 9 Oct 2023 (v1), last revised 2 Apr 2024 (this version, v2)]

Title:Rephrase, Augment, Reason: Visual Grounding of Questions for Vision-Language Models

Authors:Archiki Prasad, Elias Stengel-Eskin, Mohit Bansal
View a PDF of the paper titled Rephrase, Augment, Reason: Visual Grounding of Questions for Vision-Language Models, by Archiki Prasad and 2 other authors
View PDF HTML (experimental)
Abstract:An increasing number of vision-language tasks can be handled with little to no training, i.e., in a zero and few-shot manner, by marrying large language models (LLMs) to vision encoders, resulting in large vision-language models (LVLMs). While this has huge upsides, such as not requiring training data or custom architectures, how an input is presented to an LVLM can have a major impact on zero-shot model performance. In particular, inputs phrased in an underspecified way can result in incorrect answers due to factors like missing visual information, complex implicit reasoning, or linguistic ambiguity. Therefore, adding visually-grounded information to the input as a preemptive clarification should improve model performance by reducing underspecification, e.g., by localizing objects and disambiguating references. Similarly, in the VQA setting, changing the way questions are framed can make them easier for models to answer. To this end, we present Rephrase, Augment and Reason (RepARe), a gradient-free framework that extracts salient details about the image using the underlying LVLM as a captioner and reasoner, in order to propose modifications to the original question. We then use the LVLM's confidence over a generated answer as an unsupervised scoring function to select the rephrased question most likely to improve zero-shot performance. Focusing on three visual question answering tasks, we show that RepARe can result in a 3.85% (absolute) increase in zero-shot accuracy on VQAv2, 6.41%, and 7.94% points increase on A-OKVQA, and VizWiz respectively. Additionally, we find that using gold answers for oracle question candidate selection achieves a substantial gain in VQA accuracy by up to 14.41%. Through extensive analysis, we demonstrate that outputs from RepARe increase syntactic complexity, and effectively utilize vision-language interaction and the frozen LLM.
Comments: ICLR 2024 camera-ready (23 pages), Code: this https URL
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Cite as: arXiv:2310.05861 [cs.CL]
  (or arXiv:2310.05861v2 [cs.CL] for this version)
  https://doi.org/10.48550/arXiv.2310.05861
arXiv-issued DOI via DataCite

Submission history

From: Archiki Prasad [view email]
[v1] Mon, 9 Oct 2023 16:57:57 UTC (4,455 KB)
[v2] Tue, 2 Apr 2024 17:37:42 UTC (4,457 KB)
Full-text links:

Access Paper:

    View a PDF of the paper titled Rephrase, Augment, Reason: Visual Grounding of Questions for Vision-Language Models, by Archiki Prasad and 2 other authors
  • View PDF
  • HTML (experimental)
  • TeX Source
view license

Current browse context:

cs.CL
< prev   |   next >
new | recent | 2023-10
Change to browse by:
cs
cs.AI
cs.CV
cs.LG

References & Citations

  • NASA ADS
  • Google Scholar
  • Semantic Scholar
Loading...

BibTeX formatted citation

Data provided by:

Bookmark

BibSonomy Reddit

Bibliographic and Citation Tools

Bibliographic Explorer (What is the Explorer?)
Connected Papers (What is Connected Papers?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)

Code, Data and Media Associated with this Article

alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
ScienceCast (What is ScienceCast?)

Demos

Replicate (What is Replicate?)
Hugging Face Spaces (What is Spaces?)
TXYZ.AI (What is TXYZ.AI?)

Recommenders and Search Tools

Influence Flower (What are Influence Flowers?)
CORE Recommender (What is CORE?)
  • Author
  • Venue
  • Institution
  • Topic

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)
We gratefully acknowledge support from our major funders, member institutions, , and all contributors.
About · Help · Contact · Subscribe · Copyright · Privacy · Accessibility · Operational Status (opens in new tab)
Major funding support from
Simons Foundation Schmidt Sciences