Embodied Vision-and-Language Navigation with Dynamic Convolutional Filters
This work addresses the challenge of enabling embodied agents to navigate using natural language instructions in real-world environments, representing an incremental advance in VLN methods.
The paper tackled the problem of Vision-and-Language Navigation (VLN) by proposing a model that uses dynamic convolutional filters to encode visual and language information, achieving state-of-the-art performance in low-level action spaces with improved results over traditional convolution methods.
In Vision-and-Language Navigation (VLN), an embodied agent needs to reach a target destination with the only guidance of a natural language instruction. To explore the environment and progress towards the target location, the agent must perform a series of low-level actions, such as rotate, before stepping ahead. In this paper, we propose to exploit dynamic convolutional filters to encode the visual information and the lingual description in an efficient way. Differently from some previous works that abstract from the agent perspective and use high-level navigation spaces, we design a policy which decodes the information provided by dynamic convolution into a series of low-level, agent friendly actions. Results show that our model exploiting dynamic filters performs better than other architectures with traditional convolution, being the new state of the art for embodied VLN in the low-level action space. Additionally, we attempt to categorize recent work on VLN depending on their architectural choices and distinguish two main groups: we call them low-level actions and high-level actions models. To the best of our knowledge, we are the first to propose this analysis and categorization for VLN.