Zero-Shot and Few-Shot Video Question Answering with Multi-Modal Prompts

ICCVW 2023 pp. 2797-2802

doi:10.1109/ICCVW60793.2023.00298 /iccvw/2023/engin2023iccvw-zeroshot/

Abstract

Recent vision-language models are driven by large-scale pretrained models. However, adapting pretrained models on limited data presents challenges such as overfitting, catastrophic forgetting, and the cross-modal gap between vision and language. We introduce a parameter-efficient method to address these challenges, combining multimodal prompt learning and a transformer-based mapping network, while keeping the pretrained models frozen. Our experiments on several video question answering benchmarks demonstrate the superiority of our approach in terms of performance and parameter efficiency on both zero-shot and few-shot settings. Our code is available at https://engindeniz.github.io/vitis.

PDF ICCVW Semantic Scholar

Cite

Text

Engin and Avrithis. "Zero-Shot and Few-Shot Video Question Answering with Multi-Modal Prompts." IEEE/CVF International Conference on Computer Vision Workshops, 2023. doi:10.1109/ICCVW60793.2023.00298

Markdown

[Engin and Avrithis. "Zero-Shot and Few-Shot Video Question Answering with Multi-Modal Prompts." IEEE/CVF International Conference on Computer Vision Workshops, 2023.](https://mlanthology.org/iccvw/2023/engin2023iccvw-zeroshot/) doi:10.1109/ICCVW60793.2023.00298

BibTeX

@inproceedings{engin2023iccvw-zeroshot,
  title     = {{Zero-Shot and Few-Shot Video Question Answering with Multi-Modal Prompts}},
  author    = {Engin, Deniz and Avrithis, Yannis},
  booktitle = {IEEE/CVF International Conference on Computer Vision Workshops},
  year      = {2023},
  pages     = {2797-2802},
  doi       = {10.1109/ICCVW60793.2023.00298},
  url       = {https://mlanthology.org/iccvw/2023/engin2023iccvw-zeroshot/}
}