Reproducibility Report: Contextualizing Hate Speech Classifiers with Post-hoc Explanation

TL;DR

This report assesses the reproducibility of a hate speech classification method that uses post-hoc explanations to understand model decisions, focusing on the method's validity and experimental results.

Contribution

It provides a detailed reproducibility analysis of the proposed method and evaluates the validity of its experimental findings.

Findings

Reproducibility of the method was successfully verified.

The original results were largely confirmed.

Insights into model explanations were gained.

Abstract

The presented report evaluates Contextualizing Hate Speech Classifiers with Post-hoc Explanation paper within the scope of ML Reproducibility Challenge 2020. Our work focuses on both aspects constituting the paper: the method itself and the validity of the stated results. In the following sections, we have described the paper, related works, algorithmic frameworks, our experiments and evaluations.