Unveiling the Black Box: Explainable AI Techniques in Machine Learning

EasyChair Preprint 12241

Abstract

This paper provides an overview of the state-of-the-art techniques in XAI, focusing on their applications in machine learning. It begins by elucidating the importance of interpretability in AI systems, emphasizing the significance of trust, accountability, and ethical considerations. Subsequently, it delves into various XAI methods, categorizing them into model-specific and model-agnostic approaches. Model-specific techniques are tailored to particular types of machine learning models, such as decision trees, linear models, or neural networks. They often exploit the inherent structure or properties of these models to provide explanations. On the other hand, model-agnostic methods do not rely on specific model characteristics and can be applied universally across different types of models.

Keyphrases: across, different, universally

@booklet{EasyChair:12241,
  author    = {Kurez Oroy and Jack Nick},
  title     = {Unveiling the Black Box: Explainable AI Techniques in Machine Learning},
  howpublished = {EasyChair Preprint 12241},
  year      = {EasyChair, 2024}}