Human Inductive Biases
The discussion delves into instilling human-like inductive biases in neural network agents, emphasizing how humans quickly learn and generalize due to these biases. Through experiments with two-dimensional grids, it becomes evident that humans and machines exhibit different inductive biases, which can be bridged by training agents on natural language and program induction abstractions. This dual approach aims to enhance agents' performance on tasks where humans excel, while also ensuring they struggle with tasks that challenge human learners.In this clip
From this podcast
Machine Learning Street Talk (MLST)
#97 SREEJAN KUMAR - Human Inductive Biases in Machines from Language
Related Questions