English

Vision-and-Language Navigation via Causal Learning

Computer Vision and Pattern Recognition 2024-04-17 v1 Artificial Intelligence

Abstract

In the pursuit of robust and generalizable environment perception and language understanding, the ubiquitous challenge of dataset bias continues to plague vision-and-language navigation (VLN) agents, hindering their performance in unseen environments. This paper introduces the generalized cross-modal causal transformer (GOAT), a pioneering solution rooted in the paradigm of causal inference. By delving into both observable and unobservable confounders within vision, language, and history, we propose the back-door and front-door adjustment causal learning (BACL and FACL) modules to promote unbiased learning by comprehensively mitigating potential spurious correlations. Additionally, to capture global confounder features, we propose a cross-modal feature pooling (CFP) module supervised by contrastive learning, which is also shown to be effective in improving cross-modal representations during pre-training. Extensive experiments across multiple VLN datasets (R2R, REVERIE, RxR, and SOON) underscore the superiority of our proposed method over previous state-of-the-art approaches. Code is available at https://github.com/CrystalSixone/VLN-GOAT.

Keywords

Cite

@article{arxiv.2404.10241,
  title  = {Vision-and-Language Navigation via Causal Learning},
  author = {Liuyi Wang and Zongtao He and Ronghao Dang and Mengjiao Shen and Chengju Liu and Qijun Chen},
  journal= {arXiv preprint arXiv:2404.10241},
  year   = {2024}
}
R2 v1 2026-06-28T15:55:19.824Z