Towards Explainability for AI Fairness

Review Social Sciences, Interdisciplinary

Fairness perceptions of algorithmic decision-making: A systematic review of the empirical literature

Christopher Starke et al.

Summary: This article provides a comprehensive review of empirical studies on perceptions of algorithmic fairness. The insights come from 58 studies and cover multiple domains and scientific disciplines. Although there is heterogeneity in the theoretical concepts and empirical measurements of algorithmic fairness, the findings primarily focus on Western-democratic contexts. By advocating for interdisciplinary research and a society-in-the-loop framework, the authors aim to contribute to fairer and more responsible algorithmic decision-making.

BIG DATA & SOCIETY (2022)

Add to Collection

Review Computer Science, Information Systems

Evaluating the Quality of Machine Learning Explanations: A Survey on Methods and Metrics

Jianlong Zhou et al.

Summary: The paper provides a comprehensive overview of methods proposed for the evaluation of ML explanations in the current literature. It identifies properties of explainability from the review of definitions of explainability and uses them as objectives that evaluation metrics should achieve. The survey found that different explanation methods use quantitative metrics primarily to evaluate either simplicity of interpretability or fidelity of explainability, while subjective measures like trust and confidence are key in human-centered evaluation of explainable systems.

ELECTRONICS (2021)

Add to Collection

Article Computer Science, Artificial Intelligence

Towards multi-modal causability with Graph Neural Networks enabling information fusion for explainable AI

Andreas Holzinger et al.

Summary: AI excels in certain tasks but humans excel at multi-modal thinking and building self-explanatory systems. The medical domain highlights the importance of various modalities contributing to one result. Using conceptual knowledge to guide model training can lead to more explainable, robust, and less biased machine learning models.

INFORMATION FUSION (2021)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

Digital Transformation for Sustainable Development Goals (SDGs) - A Security, Safety and Privacy Perspective on AI

Andreas Holzinger et al.

Summary: Artificial intelligence is a major driver in the ongoing digital transformation, offering immense potential to benefit society but also posing unforeseen threats. All stakeholders must work together to ensure the development and use of AI align with human values, maintaining safety, security, and transparency in applications.

MACHINE LEARNING AND KNOWLEDGE EXTRACTION (CD-MAKE 2021) (2021)

Add to Collection

Review Computer Science, Hardware & Architecture