NeurIPS 2020
### A Statistical Mechanics Framework for Task-Agnostic Sample Design in Machine Learning

### Meta Review

The authors present a new method based on statistical mechanics for studying and comparing sample designs. The reviewers think this is an important and sound theoretical contribution. The authors should however make the changes suggested by the reviewers, for example, by adding remarks and discussions to make the theoretical results more understandable, and by clarifying differences from the existing work in [17].