Policies often fail due to distribution shift – changes in the state and...
Collaborative tasks often begin with partial task knowledge and incomple...
To act in the world, robots rely on a representation of salient task asp...
In September 2016, Stanford's "One Hundred Year Study on Artificial
Inte...
In September 2021, the "One Hundred Year Study on Artificial Intelligenc...
Emergent communication research often focuses on optimizing task-specifi...
Artificial neural nets can represent and classify many types of data but...
Feature attribution methods are popular for explaining neural network
pr...
Interpretability methods are developed to understand the working mechani...
Recent causal probing literature reveals when language models and syntac...
Neural nets are powerful function approximators, but the behavior of a g...
Communication enables agents to cooperate to achieve their goals. Learni...
Neural rationale models are popular for interpretable predictions of NLP...
Explainable AI techniques that describe agent reward functions can enhan...
Neural agents trained in reinforcement learning settings can learn to
co...
Trained AI systems and expert decision makers can make errors that are o...
As robots are deployed in complex situations, engineers and end users mu...
Robotic agents must adopt existing social conventions in order to be
eff...
Reward engineering is crucial to high performance in reinforcement learn...
Agents trained in simulation may make errors in the real world due to
mi...
Coordinating agents to complete a set of tasks with intercoupled tempora...
We present the Bayesian Case Model (BCM), a general framework for Bayesi...
We present a framework for learning human user models from joint-action
...
We aim to reduce the burden of programming and deploying autonomous syst...