Omnigram explorer: an interactive tool for understanding Bayesian networks
0 likes432 views
The document introduces Omnigram Explorer, an open-source interactive tool designed for exploring Bayesian networks, emphasizing its capabilities in visualizing complex data relationships. It features innovations like single and multi-node brushing, linked brushing, and flash mode to facilitate sensitivity analysis and the visualization of dependencies. The tool is aimed at enhancing user understanding of data sets and interaction with Bayesian networks, despite not being fully integrated with a Bayesian network API yet.
Omnigram explorer: an interactive tool for understanding Bayesian networks
- 1. 1 Omnigram Explorer: A New Interactive Tool for Exploring Bayesian Networks Kevin B Korb1 & Tim Taylor2 1Faculty of Information Technology, Monash University 2University of London International Programmes www.tim-taylor.com/omnigram Australasian Bayesian Network Modelling Society 2015
- 2. Omnigram Explorer 1. Motivation 2. Omnigram Explorer Tools a. Concepts b. Single Node Brushing c. Multi Node Brushing d. Linked Brushing e. Flash Mode 3. Exploring Dependencies a. Viewing D-Separation b. Explaining Away c. The Best Explanation 4. Conclusion
- 3. Motivation Google’s Ngram Viewer reports the relentless growth of Bayesian networks: § Parallel increases in complexity & data. § New ways of interactively visualizing both networks & data are needed.
- 4. Omnigram Explorer § Developed by Tim Taylor; open source – www.tim-taylor.com – https://github.com/tim-taylor/omnigram § Traditional visualization of multiple variables: – Scatterplots – Parallel Coordinates: allow tracing samples across many dimensions
- 5. Omnigram Explorer Two main innovations: § Visualization takes advantage of human visual system, especially psychology of pattern and motion perception § Interactive controls allow user to manipulate patterns, visualizing not just static data but active dependencies between variables In particular, OE enables interactive sensitivity analysis of Bayesian networks with user-chosen sets of observations (“sensitivity to data”)
- 6. OE OE Window BN based on UCI car+mpg data set
- 7. OE: Single Node Brushing § “1” starts single node brushing § Enter observation as a range in one variable; observe consequences § Drag range to observe dependencies
- 8. OE: Multi Node Brushing § “2” starts multi node brushing § Enter observations as ranges in multiple variables; colors show full & partial matches § Drag ranges to observe dependencies
- 9. OE: Linked Brushing § In multi node brushing, hit “L” with cursor over successive nodes § Drag range in either variable, linked variables follow § You can reverse link direction (negative dependency); link’s box will invert
- 10. OE: Flash Mode § “3” starts flash mode § Cycles through small (or large) samples, replacing oldest with new § Speed and sample size controllable
- 11. Exploring Dependencies: D-Separation § In multi node, fix an observation set (here cylinders & horsepower) § Observe which variables respond when varying another (e.g., model year) § Partial observations will “leak”; other leakage indicates (some) violation of the Markov property
- 12. Exploring Dependencies: Explaining Away § In multi node, fix a child of more than one parent; explore the induced dependency between parents (“explaining away”) § Here early model year partially explains high displacement, so Origin=1 (US) declines
- 13. Exploring Dependencies: Inference to the Best Explanation § What could explain an unusual event? E.g., high acceleration + good mpg? § Read off the means for precursor variables: late model, foreign, 4 cylinders, low weight, etc.
- 14. Conclusion OE is a great interactive tool for either: § Exploring a data set, to get a feel for dependencies and independencies, relations worth exploring § Exploring a Bayesian network: – To get a feel for dependencies and independencies, relations worth exploring – Examine d-separation properties – Performing sensitivity analysis interactively & visually NB: OE is not (yet) implemented via a BN API, but can be used via sampling with any BN.
- 15. References § Taylor, T., Dorin, A., & Korb, K. Omnigram Explorer: A Simple Interactive Tool for the Initial Exploration of Complex Systems. European Conference on Artificial Life, 2015. § Ropero, R. F., Nicholson, A. E., & Korb, K. (2015). Using a New Tool to Visualize Environmental Data for Bayesian Network Modelling. In Advances in Artificial Intelligence (pp. 175-184). Springer International Publishing.