There has been an emerging paradigm shift from the era of “internet AI” to “embodied AI,”
where AI algorithms and agents no longer learn from datasets of images, videos or text …

Learning to generate diverse scene-aware and goal-oriented human motions in 3D scenes
remains challenging due to the mediocre characters of the existing datasets on Human …

Moving around in the world is naturally a multisensory experience, but today's embodied
agents are deaf—restricted to solely their visual perception of the environment. We introduce …

For robots to be generally useful, they must be able to find arbitrary objects described by
people (ie, be language-driven) even without expensive navigation training on in-domain …

With the recent development of deep learning technology comes the wide use of artificial
intelligence (AI) models in various domains. AI shows good performance for definite …

Reinforcement Learning (RL) aims at learning an optimal behavior policy from its own
experiments and not rule-based control methods. However, there is no RL algorithm yet …

Does progress in simulation translate to progress on robots? If one method outperforms
another in simulation, how likely is that trend to hold in reality on a robot? We examine this …

In reinforcement learning for visual navigation, it is common to develop a model for each
new task, and train that model from scratch with task-specific interactions in 3D …

Navigation tasks in photorealistic 3D environments are challenging because they require
perception and effective planning under partial observability. Recent work shows that map …

Can general-purpose neural models learn to navigate? For PointGoal navigation ("" go to x,
y""), the answer is a clearyes'--mapless neural models composed of task-agnostic …