What is amortization and equalization

[LANSHANGH]

I am very interested in your task, but I find that I do not understand your paper. Is there any book or article that can help me understand what problem you solve.

[iancovert]

The paper is about certain tasks in explainable machine learning that perform expensive computation for each input (e.g., data valuation scores that must be calculated for each labeled pair in the dataset), and how to make them faster using amortization.

To understand the explainable machine learning tasks we focus on, you could read this paper which covers many feature attribution methods, and this paper about data valuation.

To understand amortization, you could read this paper which provides a broad overview of this framework and applications across machine learning. The paper is quite detailed so reading through section 2.2.1 should be sufficient to understand our paper.