Work

Exploring the Plausibility of Hate and Counter Speech Detectors with Explainable AI

Year: 2024

Type: conference-paper

Abstract: In this paper we investigate the explainability of transformer models and their plausibility for hate speech and counter speech detection. We compare representatives of four different explainability approaches, i.e., gradient-based, perturbation-based, attention-based, and prototype-based approaches, and analyze them quantitatively with an ablation study and qualitatively in a user study. Results show that perturbationbased explainability performs best, followed by gradient-based and attention-based explainability. Prototype-based experiments did not yield useful results. Overall, we observe that explainability strongly supports the users in better understanding the model predictions.

Source: International Conference on Content-Based Multimedia Indexing (CBMI)

DOI: 10.1109/cbmi62980.2024.10859247

Authors: Adrian Jaques Böck, Djordje Slijepčević, Matthias Zeppelzauer

Institution: University of Applied Sciences St Pölten

Language: English

Topic: Explainable Artificial Intelligence (XAI)

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

SDG: Peace, justice, and strong institutions

Open Access status: green

Funders: Vienna Science and Technology Fund, Österreichische Forschungsförderungsgesellschaft