A Framework for Improving Scholarly Neural Network Diagrams

• Artificial intelligence system experts are interviewed about their use of diagrams. • Differences in interpretation, preference and use of scholarly system diagrams are discovered. • Priorities and problems that scholarly system diagram users encounter are identified. • A guidelines-based framework for neural network system diagrams is proposed. • The proposed framework is evaluated in a multi-stage mixed-methods empirical study, and with a quantitative corpus-based approach.