https://arxiv.org/abs/2302.00923

close this message
arXiv smileybones icon

Global Survey

In just 3 minutes help us understand how you see arXiv.

TAKE SURVEY
Skip to main content
Cornell University
We gratefully acknowledge support from
the Simons Foundation and member institutions.
 
arxiv logo > cs > arXiv:2302.00923
[                    ]

Help | Advanced Search

[All fields        ]
Search
arXiv logo
Cornell University Logo
[                    ] GO
quick links

  * Login
  * Help Pages
  * About

Computer Science > Computation and Language

arXiv:2302.00923 (cs)
[Submitted on 2 Feb 2023]

Title:Multimodal Chain-of-Thought Reasoning in Language Models

Authors:Zhuosheng Zhang, Aston Zhang, Mu Li, Hai Zhao, George Karypis
, Alex Smola
Download PDF

    Abstract: Large language models (LLMs) have shown impressive
    performance on complex reasoning by leveraging chain-of-thought
    (CoT) prompting to generate intermediate reasoning chains as the
    rationale to infer the answer. However, existing CoT studies are
    mostly isolated in the language modality with LLMs, where LLMs
    are hard to deploy. To elicit CoT reasoning in multimodality, a
    possible solution is to fine-tune small language models by fusing
    the vision and language features to perform CoT reasoning. The
    key challenge is that those language models tend to generate
    hallucinated reasoning chains that mislead the answer inference.
    To mitigate the effect of such mistakes, we propose
    Multimodal-CoT that incorporates vision features in a decoupled
    training framework. The framework separates the rationale
    generation and answer inference into two stages. By incorporating
    the vision features in both stages, the model is able to generate
    effective rationales that contribute to answer inference. With
    Multimodal-CoT, our model under 1 billion parameters outperforms
    the previous state-of-the-art LLM (GPT-3.5) by 16% (75.17%->
    91.68%) on the ScienceQA benchmark and even surpasses human
    performance. Code is publicly available at this https URL.

Subjects: Computation and Language (cs.CL); Artificial Intelligence
          (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
Cite as:  arXiv:2302.00923 [cs.CL]
          (or arXiv:2302.00923v1 [cs.CL] for this version)
          https://doi.org/10.48550/arXiv.2302.00923
          Focus to learn more
          arXiv-issued DOI via DataCite

Submission history

From: Aston Zhang [view email]
[v1] Thu, 2 Feb 2023 07:51:19 UTC (421 KB)
Full-text links:

Download:

  * PDF
  * Other formats

[by-sa-4]
Current browse context:
cs.CL
< prev   |   next >
new | recent | 2302
Change to browse by:
cs
cs.AI
cs.CV

References & Citations

  * NASA ADS
  * Google Scholar
  * Semantic Scholar

a export bibtex citation Loading...

Bibtex formatted citation

x
[loading...          ]
Data provided by:

Bookmark

BibSonomy logo Mendeley logo Reddit logo ScienceWISE logo
(*) Bibliographic Tools

Bibliographic and Citation Tools

[ ] Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
[ ] Litmaps Toggle
Litmaps (What is Litmaps?)
[ ] scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
( ) Code, Data, Media

Code, Data and Media Associated with this Article

[ ] Links to Code Toggle
Papers with Code (What is Papers with Code?)
[ ] ScienceCast Toggle
ScienceCast (What is ScienceCast?)
( ) Demos

Demos

[ ] Replicate Toggle
Replicate (What is Replicate?)
[ ] Spaces Toggle
Hugging Face Spaces (What is Spaces?)
( ) Related Papers

Recommenders and Search Tools

[ ] Connected Papers Toggle
Connected Papers (What is Connected Papers?)
[ ] Core recommender toggle
CORE Recommender (What is CORE?)
( ) About arXivLabs

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?)

  * About
  * Help

  * Click here to contact arXiv Contact
  * Click here to subscribe Subscribe

  * Copyright
  * Privacy Policy

  * Web Accessibility Assistance
  * arXiv Operational Status
    Get status notifications via email or slack