Abstract
In this paper, we propose an error causal inference method that could be used for finding dominant features for a faulty instance under a well-trained multi-modality input model, which could apply to any testing instance. We evaluate our method using a well-trained multi-modalities stylish caption generation model and find those causal inferences that could provide us the insights for next step optimization.- Anthology ID:
- 2021.alvr-1.2
- Volume:
- Proceedings of the Second Workshop on Advances in Language and Vision Research
- Month:
- June
- Year:
- 2021
- Address:
- Online
- Venue:
- ALVR
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 11–15
- Language:
- URL:
- https://aclanthology.org/2021.alvr-1.2
- DOI:
- 10.18653/v1/2021.alvr-1.2
- Cite (ACL):
- Chengxi Li and Brent Harrison. 2021. Error Causal inference for Multi-Fusion models. In Proceedings of the Second Workshop on Advances in Language and Vision Research, pages 11–15, Online. Association for Computational Linguistics.
- Cite (Informal):
- Error Causal inference for Multi-Fusion models (Li & Harrison, ALVR 2021)
- PDF:
- https://preview.aclanthology.org/auto-file-uploads/2021.alvr-1.2.pdf