Research Results | Improving IV&V Techniques Through the Analysis of Anomalies


	Search:	Advanced Search

Click here to complete a short survey. The results of this survey will be used to help us improve the research program and this website.

Click here to view research projects that had new research results added in the last 90 days.

Improving IV&V Techniques Through the Analysis of Anomalies
Point of Contact	Tim Menzies tim@menzies.us
Dates	June 2006 - December 2008
Problem	Generating a catalogue of best practices for software IV&V is complicated by the diverse nature of NASA software development as well as the complexities of record keeping for an organization as diverse as NASA. Here, AI tools will be used to augment NASA?s databases of IV&V events these tools will allow us to infer what is currently hidden: i.e. what are the demonstrably superior best practices?
Objective	# Q: How to recognize an anomalous project? # A1: (Year1+2) Automatic analysis of NASA IV&V databases. SILAP, etc * But? Database structures at NASA IV&V keep changing # A2: (Year 3) * Build a fast-change format for intuitions about a project * Use manager expectations for those intuitions * Initialize the intuitions using Dabney?s Bayes nets * Build a GUI to make the change easy * Conduct user trials
Results	SAS 06 Executive Presentation.pdf SAS 06 Technical Presentation.pdf SAS_07_Exec_Brief_Analysis_Anomalies_Menzies.ppt SAS_07_Tech_Pres_Analysis_Anomalies_Menzies.pdf SAS 2008 Executive Presentation - Improving IV&V Techniques through the Analysis of Anomalies.pdf
Keywords	anomaly, detection, artificial intelligence, verification, validation, Bayes
Categories	Code Analysis Design Analysis Domain-Specific Analysis Dynamic Analysis Management & Planning of V&V Process Improvement Static Analysis

Curator: Josh Stonestreet
NASA Official: Lisa Montgomery