To interact with humans in the world, agents need to understand the dive...
Reinforcement learning from human feedback (RLHF) is a technique for tra...
Our goal is for robots to follow natural language instructions like "put...
Consider a robot tasked with tidying a desk with a meticulously construc...
Deep reinforcement learning (RL) works impressively in some environments...
Auditing large language models for unexpected behaviors is critical to
p...
In the real world, some of the most complex settings for learned agents
...
Humans have internal models of robots (like their physical capabilities)...
Inferring reward functions from human behavior is at the center of value...
One of the most successful paradigms for reward learning uses human feed...
Randomly masking and predicting word tokens has been a successful approa...
AI agents designed to collaborate with people benefit from models that e...
Randomly masking and predicting word tokens has been a successful approa...
The content that a recommender system (RS) shows to users influences the...
Models of human behavior for prediction and collaboration tend to fall i...
In classic instruction following, language like "I'd like the JetBlue fl...
Assuming humans are (approximately) rational enables robots to infer rew...
Reinforcement learning (RL) requires access to a reward function that
in...
The last decade has seen a significant increase of interest in deep lear...
Since reward functions are hard to specify, recent work has focused on
l...
Specifying reward functions for robots that operate in environments with...
One difficulty in using artificial agents for human-assistive applicatio...
While we would like agents that can coordinate with humans, current
algo...
Reinforcement learning (RL) agents optimize only the features specified ...
Learning preferences implicit in the choices humans make is a well studi...
Fundamental to robotics is the debate between model-based and model-free...
A significant challenge for the practical application of reinforcement
l...
In shared autonomy, user input is combined with semi-autonomous control ...
Autonomous agents optimize the reward function we give them. What they d...
Intuitively, obedience -- following the order that a human gives -- seem...
Several approaches have recently been proposed for learning decentralize...
It is clear that one of the primary tools we can use to mitigate the
pot...
For an autonomous system to be helpful to humans and to pose no unwarran...