Asking 'Why' in AI: Explainability of intelligent systems - perspectives and challenges

By Alun David Preece. 2018

145 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Peeking Inside the Black-Box: A Survey on Explainable Artificial Intelligence (XAI) (2018, background): “How exactly to combine elements from both classical explainable expert systems and present interpretable ML methods is a topic of debate, for instance, Preece [57] argues that elements of that earlier work on expert systems offer routes to making progress towards XAI today.”
Explainable AI (XAI): A Systematic Meta-Survey of Current Challenges and Future Opportunities (2021, background): “Deployment Improving explanations with ontologies [18] XAI and privacy [35, 58] XAI and security [2, 64, 62, 30] XAI and safety [65, 66, 2] Human-machine teaming [40, 30, 28, 66, 1, 50, 59, 24], [45, 46]* Explainable agency [67] Machine-to-machine explanation [29, 68, 1] XAI and reinforcement learni”
Artificial Intelligence (AI): Multidisciplinary perspectives on emerging challenges, opportunities, and agenda for research, practice and policy (2019, background): “Explanation approaches are typically divided into transparency and post hoc interpretation (see e.g. Preece, 2018), the former being more suitable for “ how ” explanations, the latter for “ why ” .”
Explainable Artificial Intelligence: a Systematic Review (2020, methodology): “Application-grounded metrics assess the quality of machineproduced explanations of data-driven models by comparing the increase in productivity of a few users of these models when following these explanations instead of those produced by human engineers, as done in [26, 2].”
The Ethical Implications of Using Artificial Intelligence in Auditing (2020, background): “However, the algorithms that form the backbone of AI may evolve and adopt such complex workings rendering them difficult to understand for even their developers (Preece 2018).”