A Recurrent Vision-and-Language BERT for Navigation
This work provides a more efficient and generalizable transformer-based solution for vision-and-language navigation tasks, benefiting researchers and developers working on embodied AI agents.
This paper proposes a recurrent BERT model for vision-and-language navigation (VLN) that addresses the challenge of history-dependent attention and decision-making in partially observable Markov decision processes. The model achieves state-of-the-art results on R2R and REVERIE datasets, outperforming more complex encoder-decoder models.
Accuracy of many visiolinguistic tasks has benefited significantly from the application of vision-and-language(V&L) BERT. However, its application for the task of vision-and-language navigation (VLN) remains limited. One reason for this is the difficulty adapting the BERT architecture to the partially observable Markov decision process present in VLN, requiring history-dependent attention and decision making. In this paper we propose a recurrent BERT model that is time-aware for use in VLN. Specifically, we equip the BERT model with a recurrent function that maintains cross-modal state information for the agent. Through extensive experiments on R2R and REVERIE we demonstrate that our model can replace more complex encoder-decoder models to achieve state-of-the-art results. Moreover, our approach can be generalised to other transformer-based architectures, supports pre-training, and is capable of solving navigation and referring expression tasks simultaneously.