Images in Language Space: Exploring the Suitability of Large Language Models for Vision & Language Tasks

Publication

This is the official Git repository page for the paper:

Hakimov, S., and Schlangen, D., (2023). Images in Language Space: Exploring the Suitability of Large Language Models for Vision & Language Tasks. Findings of the Association for Computational Linguistics: ACL 2023 PDF

Code

Installation

Create a virtual environment in Python 3 and run the following scripts

pip install -r requirements.txt

pip install git+https://github.com/openai/CLIP.git

To use OFA Captioning model, please check the instructions from their official Git repository.

Extraction of Image as Text Representation

Run the following code snippet that extracts visual tags, captions from BLIP and VITGPT2 models for images in the 'images' directory

python extract_image_as_text_representation.py --dir images

Prompting Large Language Models

The prompt templates for each dataset are given in the resource/prompt_templates directory.

Citation

If you find the resources or the code useful, please cite us:

@inproceedings{hakimov-schlangen-2023-images,
    title = "Images in Language Space: Exploring the Suitability of Large Language Models for Vision {\&} Language Tasks",
    author = "Hakimov, Sherzod  and
      Schlangen, David",
    booktitle = "Findings of the Association for Computational Linguistics: ACL 2023",
    month = jul,
    year = "2023",
    address = "Toronto, Canada",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2023.findings-acl.894",
    pages = "14196--14210",
    abstract = "Large language models have demonstrated robust performance on various language tasks using zero-shot or few-shot learning paradigms. While being actively researched, multimodal models that can additionally handle images as input have yet to catch up in size and generality with language-only models. In this work, we ask whether language-only models can be utilised for tasks that require visual input {--} but also, as we argue, often require a strong reasoning component. Similar to some recent related work, we make visual information accessible to the language model using separate verbalisation models. Specifically, we investigate the performance of open-source, open-access language models against GPT-3 on five vision-language tasks when given textually-encoded visual information. Our results suggest that language models are effective for solving vision-language tasks even with limited samples. This approach also enhances the interpretability of a model{'}s output by providing a means of tracing the output back through the verbalised image content.",
}