Causality-based Cross-Modal Representation Learning for Vision-and-Language Navigation

arXiv:2403.03405v1 Announce Type: new
Abstract: Vision-and-Language Navigation (VLN) has gained significant research interest in recent years due to its potential applications in real-world scenarios. However, existing VLN methods struggle with the issue of spurious associations, resulting in poor generalization with a significant performance gap between seen and unseen environments. In this paper, we tackle this challenge by proposing a unified framework CausalVLN based on the causal learning paradigm to train a robust navigator capable of learning unbiased feature representations. Specifically, …

abstract applications arxiv causality cs.cv environments gap however issue language modal navigation paper performance representation representation learning research struggle type vision vision-and-language world