Learning to see and act

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Towards Causal Representation Learning (2021, background): “In these successes, there are several trends at work [215]: (1) we have massive amounts of data, often from simulations or large scale human labeling, (2) we use high capacity machine learning systems (i.”
Causality for Machine Learning (2019, background): “Two major issues can be stated in terms of questions (Schölkopf, 2015, 2017): Question 1: why is RL on the original high-dimensional ATARI games harder than on downsampled versions?”
Inductive biases for deep learning of higher-level cognition (2020, methodology): “We use the term structural-mechanistic (Schölkopf, 2015) to characterize models which follow an underlying mechanistic understanding of reality.”
A Hierarchy of Limitations in Machine Learning (2020, background): “Bernhard Schölkopf (2015) references old behavioralist language (Katz, 2017) in analogizing this to “superstitious behavior,” “in which statistical associations may be misinterpreted as causal” (although I would say that just because a model uses correlation for prediction, it does not mean we…”