Oct 30, 2022 · We propose Vienna, a versatile embodied navigation agent that simultaneously learns to perform the four navigation tasks with one model.

In response, a large-scale 3D dataset, VXN, is established to investigate multitask multimodal embodied navigation in audiovisual complex indoor environments.

The official repository of the NeurIPS 2022 paper: Towards Versatile Embodied Navigation. VXN is the abbreviation of Vision-$X$ Navigation.

Multitask Distributed Proximal Policy Optimization. We propose MDPPO based on PPO [1] that is an on-policy RL algorithm with actor-critic architecture.

First, we propose VXN, a large-scale 3D dataset that instantiates four classic navigation tasks in standardized, continuous, and audiovisual-rich environments.

Building a generalist agent that can interact with the world is the intriguing target of AI systems, thus spurring the research for embodied navigation, ...

Explore all code implementations available for Towards Versatile Embodied Navigation.

Pull requests. Repository of our accepted NeurIPS-2022 paper "Towards Versatile Embodied Navigation". embodied-navigation neurips-2022. Updated on Dec 8, 2022 ...

Visual Navigation is the problem of navigating an agent, e.g. a mobile robot, in an environment using camera input only. The agent is given a target image ...

Drawing on various disciplines such as neuroscience, psychology, robotics, and artificial intelligence, Embodied Intelligence seeks to create innovative models ...