Back to Search
Start Over
ET tu, CLIP? Addressing Common Object Errors for Unseen Environments
- Source :
- Conference on Computer Vision and Pattern Recognition (CVPR 2022) - Embodied AI Workshop
- Publication Year :
- 2024
-
Abstract
- We introduce a simple method that employs pre-trained CLIP encoders to enhance model generalization in the ALFRED task. In contrast to previous literature where CLIP replaces the visual encoder, we suggest using CLIP as an additional module through an auxiliary object detection objective. We validate our method on the recently proposed Episodic Transformer architecture and demonstrate that incorporating CLIP improves task performance on the unseen validation set. Additionally, our analysis results support that CLIP especially helps with leveraging object descriptions, detecting small objects, and interpreting rare words.
Details
- Database :
- arXiv
- Journal :
- Conference on Computer Vision and Pattern Recognition (CVPR 2022) - Embodied AI Workshop
- Publication Type :
- Report
- Accession number :
- edsarx.2406.17876
- Document Type :
- Working Paper