📰 News

• 30/08/2023: The preprint of our paper is now available on arXiv.

Summary

• 📰 News
• Summary
• 📚 Introduction
• 📊 Baselines
• Citation
• References

📚 Introduction

AffectVisDial is a large-scale dataset which consists of 50K 10-turn visually grounded dialogs as well as concluding emotion attributions and dialog-informed textual emotion explanations.

📊 Baselines

We provide baseline models explanation generation task:

• GenLM: BERT- and BART-based models [3, 4]
• NLX-GPT: NLX-GPT based model [1]

Citation

If you use our dataset, please cite the two following references:

@article{haydarov2023affective,
  title={Affective Visual Dialog: A Large-Scale Benchmark for Emotional Reasoning Based on Visually Grounded Conversations},
  author={Haydarov, Kilichbek and Shen, Xiaoqian and Madasu, Avinash and Salem, Mahmoud and Li, Li-Jia and Elsayed, Gamaleldin and Elhoseiny, Mohamed},
  journal={arXiv preprint arXiv:2308.16349},
  year={2023}
}

References

1. _[Sammani et al., 2022] - NLX-GPT: A Model for Natural Language Explanations in Vision and Vision-Language Tasks
2. [Li et al., 2022] - BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation
3. [Lewis et al., 2019] - BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension.
4. [Dewlin et al., 2018] - BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding