Explainable Deep Learning: A Visual Analytics Approach with Transition Matrices
Mathematics, Vol. 12, No. 7, Article 1024
Editorial summary
A transition matrix maps deep-network features into a more interpretable model for inspection through visual analytics. Human review is part of the explanation process. Experiments on MNIST, FNC-1 and Iris compare the resulting representations and classification outputs with reference data.
Bibliographic reference
Use the publisher’s record for the citation format required by your journal or organisation.
Pavlo Radiuk, Oleksander Barmak, Eduard Manziuk, Iurii Krak. "Explainable Deep Learning: A Visual Analytics Approach with Transition Matrices". Mathematics, Vol. 12, No. 7, Article 1024, 2024. https://doi.org/10.3390/math12071024
Test a method against your business task.
Describe the decision your system needs to support and the data you have. We assess whether a method fits, define a prototype and agree how to measure its quality, response time and operating cost.