We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CV

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computer Vision and Pattern Recognition

Title: Perceive, Transform, and Act: Multi-Modal Attention Networks for Vision-and-Language Navigation

Abstract: Vision-and-Language Navigation (VLN) is a challenging task in which an agent needs to follow a language-specified path to reach a target destination. In this paper, we strive for the creation of an agent able to tackle three key issues: multi-modality, long-term dependencies, and adaptability towards different locomotive settings. To that end, we devise "Perceive, Transform, and Act" (PTA): a fully-attentive VLN architecture that leaves the recurrent approach behind and the first Transformer-like architecture incorporating three different modalities - natural language, images, and discrete actions for the agent control. In particular, we adopt an early fusion strategy to merge lingual and visual information efficiently in our encoder. We then propose to refine the decoding phase with a late fusion extension between the agent's history of actions and the perception modalities. We experimentally validate our model on two datasets and two different action settings. PTA surpasses previous state-of-the-art architectures for low-level VLN on R2R and achieves the first place for both setups in the recently proposed R4R benchmark. Our code is publicly available at this https URL
Subjects: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL); Machine Learning (cs.LG)
Cite as: arXiv:1911.12377 [cs.CV]
  (or arXiv:1911.12377v1 [cs.CV] for this version)

Submission history

From: Federico Landi [view email]
[v1] Wed, 27 Nov 2019 19:00:24 GMT (747kb,D)

Link back to: arXiv, form interface, contact.