This repository contains the code for the algorithm Instance Credibility Inference (ICI).
In V1, we implement an algorithm for measuring the credibility of pseudo-labeled instances for few-shot learning by a linear regression model with incidental parameters.
In V2, we extend our algorithm to a logistic regression model with incidental parameters.
Please read the corresponding readme under each folder for a detailed introduction.