Situated Mapping of Sequential Instructions to Actions with Single-step Reward Observation.

Alane Suhr Yoav Artzi

Published in: ACL (1) (2018)

Keyphrases

single step
multi step
reward function
reinforcement learning
internal state
goal directed
decision trees
situation calculus
decision theoretic
plan recognition
learning algorithm
feature extraction
objective function
loss function
distributed cognition
initially unknown