Youtor
Understanding Argumentative Explainable AI
>> YOUR LINK HERE: ___ http://youtube.com/watch?v=IbvQc7Nsytw
Welcome to our video on Argumentative Explainable AI (XAI), presented by the Computational Logic and Argumentation Group at Imperial College London. This research aims to enhance the transparency of AI decisions, employing techniques from computational argumentation to clarify machine learning model outputs. • Argumentation frameworks form the foundation of argumentative XAI. These enable us to abstract complex AI systems in terms of understandable models from which argumentative explanations can be drawn. Argumentation frameworks can also form the core of intrinsically interpretable models for various settings, such as causal discovery and clinical decision-making. • We generate argumentative explanations using methods like dispute trees and attribution explanations, allowing both AI systems and humans to engage in meaningful argumentative exchanges. This interactive process helps achieve consensus and improves decision-making. • Presenters • Francesca Toni • Adam Dejl • Gabriel Freedman • Fabrizio Russo • Adam Gould • Guilherme Paulino-Passos • Xiang Yin • Thank you for watching, and for more details, please explore our publications. • https://www.imperial.ac.uk/computatio... • Acknowledgements • We thank Ahmed Abdullahi Idle, who filmed and edited the video, Brython Caley-Davies, who supported the projection of the slides underpinning the video on the Data Observatory at the Data Science Institute, Imperial College London, and the Data Science Institute for hosting the recording of the video. • We also thank funding for the research underpinning the video, from ERC under the EU’s Horizon 2020 research and innovation programme (grant agreement No. 101020934), from J.P. Morgan and the Royal Academy of Engineering under the Research Chairs and Senior Research Fellowships scheme and from UKRI through the CDTs in Safe and Trusted Artificial Intelligence and in AI for Health. • Finally, we thank Antonio Rago and Hamed Ayoobi for contributing some of the presented material and the other members of the CLArg group at Imperial College London for their feedback, specifically Anna Rapberger, Francesco Leofante, Deniz Gorur, Avinash Kori, and Lihu Chen.
#############################
