Visualizing Dataflow Graphs of Deep Learning Models in TensorFlow
- 29 August 2017
- journal article
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Visualization and Computer Graphics
- Vol. 24 (1), 1-12
- https://doi.org/10.1109/tvcg.2017.2744878
Abstract
We present a design study of the TensorFlow Graph Visualizer, part of the TensorFlow machine intelligence platform. This tool helps users understand complex machine learning architectures by visualizing their underlying dataflow graphs. The tool works by applying a series of graph transformations that enable standard layout techniques to produce a legible interactive diagram. To declutter the graph, we decouple non-critical nodes from the layout. To provide an overview, we build a clustered graph using the hierarchical structure annotated in the source code. To support exploration of nested structure on demand, we perform edge bundling to enable stable and responsive cluster expansion. Finally, we detect and highlight repeated structures to emphasize a models modular composition. To demonstrate the utility of the visualizer, we describe example usage scenarios and report user feedback. Overall, users find the visualizer useful for understanding, debugging, and sharing the structures of their models.This publication has 33 references indexed in Scilit:
- The Readability of Path‐Preserving Clusterings of GraphsComputer Graphics Forum, 2010
- 5 Demand Characteristics and the Concept of Quasi-ControlsPublished by Oxford University Press (OUP) ,2009
- Constrained graph layout by stress majorization and gradient projectionDiscrete Mathematics, 2009
- Centrality Based Visualization of Small World GraphsComputer Graphics Forum, 2008
- Topological Fisheye Views for Visualizing Large GraphsIEEE Transactions on Visualization and Computer Graphics, 2005
- Applying Crossing Reduction Strategies to Layered Compound GraphsLecture Notes in Computer Science, 2002
- Gradient-based learning applied to document recognitionProceedings of the IEEE, 1998
- Layout Adjustment and the Mental MapJournal of Visual Languages & Computing, 1995
- A technique for drawing directed graphsIEEE Transactions on Software Engineering, 1993
- Visualization of structural information: automatic drawing of compound digraphsIEEE Transactions on Systems, Man, and Cybernetics, 1991