You are here

ViA-ML: A Machine Learning backed Visualization Assistant

Award Information
Agency: Department of Defense
Branch: Office of the Secretary of Defense
Contract: FA8650-14-M-6532
Agency Tracking Number: O133-LD1-1211
Amount: $147,873.00
Phase: Phase I
Program: SBIR
Solicitation Topic Code: OSD13-LD1
Solicitation Number: 2013.3
Timeline
Solicitation Year: 2013
Award Year: 2014
Award Start Date (Proposal Award Date): 2014-03-18
Award End Date (Contract End Date): 2014-09-18
Small Business Information
500 West Cummings Park - Ste 3000
Woburn, MA 01801-6562
United States
DUNS: 859244204
HUBZone Owned: No
Woman Owned: No
Socially and Economically Disadvantaged: Yes
Principal Investigator
 Avinash Gandhe
 Senior Research Engineer
 (781) 933-5355
 avinash.gandhe@ssci.com
Business Contact
 Jay Miselis
Title: Director of Finance
Phone: (781) 933-5355
Email: contracts@ssci.com
Research Institution
N/A
Abstract

With the proliferation of cheap sensors, reduction in storage costs and the ubiquity of communication networks, Cyber-Physical Systems are collecting and storing data at an unprecedented rate. Analysis of such large databases is necessary to find relevant information and improve the efficiency of the Cyber Physical System. The goal of an analysis tool, simply put is to find the most interesting information in the data and present it to the user in most intuitive and clear manner possible by effectively mapping the information to visual cues. In response to this need, SSCI is proposing the development of ViA-ML, a visualization assistant with an interest-driven machine learning back-end to allow users to interactively extract information from large datasets collected by cyber physical systems. Our proposed approach is based on providing analysts with visualizations that maximize view comprehension, using pyschophysics based criteria, of the raw data attributes and attributes derived from automated analysis. Maximizing viewer comprehension then allows us to quickly gauge user interest, iterate through competing hypotheses by our novel machine learning algorithms and further enhance the visualizations by incorporating user knowledge and requirements.

* Information listed above is at the time of submission. *

US Flag An Official Website of the United States Government