Airbert: In-domain Pretraining for Vision-and-Language Navigation

TL;DR

Airbert introduces a large-scale in-domain dataset and pretraining method for vision-and-language navigation, significantly improving generalization and performance on multiple benchmarks, especially in few-shot scenarios.

Contribution

The paper presents BnB, a new large-scale in-domain VLN dataset and a pretraining approach that enhances VLN agent generalization and performance.

Findings

Outperforms state-of-the-art on R2R and REVERIE benchmarks.

Significantly improves few-shot VLN performance.

Introduces a shuffling loss for better temporal order learning.

Abstract

Vision-and-language navigation (VLN) aims to enable embodied agents to navigate in realistic environments using natural language instructions. Given the scarcity of domain-specific training data and the high diversity of image and language inputs, the generalization of VLN agents to unseen environments remains challenging. Recent methods explore pretraining to improve generalization, however, the use of generic image-caption datasets or existing small-scale VLN environments is suboptimal and results in limited improvements. In this work, we introduce BnB, a large-scale and diverse in-domain VLN dataset. We first collect image-caption (IC) pairs from hundreds of thousands of listings from online rental marketplaces. Using IC pairs we next propose automatic strategies to generate millions of VLN path-instruction (PI) pairs. We further propose a shuffling loss that improves the learning of temporal order inside PI pairs. We use BnB pretrain our Airbert model that can be adapted to discriminative and generative settings and show that it outperforms state of the art for Room-to-Room (R2R) navigation and Remote Referring Expression (REVERIE) benchmarks. Moreover, our in-domain pretraining significantly increases performance on a challenging few-shot VLN evaluation, where we train the model only on VLN instructions from a few houses.