George Mason University

Innovation House was a study funded by the Defense Advanced Research Projects Agency (DARPA) and administered by George Mason University in Fall 2012. The DARPA Innovation House looked into the feasibility of effective software design and development in a short-fuse, crucible-style living and working environment. DARPA selected imagery analysis as the topic for the effort. DARPA aimed to show that small teams of highly focused, collaborative developers operating under extremely short deadlines can make breakthroughs in automatically obtaining meaning from photos, videos, geospatial data and other imagery-related data.

The primary goal of this study was to test the feasibility of the approach by encouraging teams to explore non-traditional, novel approaches without fear of failure. Because of the high-risk nature of the desired proposal approaches,

DARPA recognized that not every attempt would succeed as envisioned. Failure of the original attempt was acceptable as long as significant learning was demonstrated as to why the radical approach failed, and recommendations were provided for alternative, follow-on approaches that could benefit the imagery analysis community.

The project was also split into two four-week phases, with second phase funding dependent on success with the first phase assignment — specifically, teams were required to “produce an initial design and demonstrate in software the crucial capabilities that validate their approach.” In phase two, teams, essentially, demonstrated a proof-of-concept software configuration. The project, however, was meant to be collaborative — not competitive.

Except for the requirement to produce software capabilities that obtain understanding from visual, geospatial and imagery-related data, there were no limitations on the applications, products, systems, modules, algorithms or any other software-realizable capabilities that might be proposed. They were to be run on any electronic systems, including but not limited to computers, tablets and mobile phones.

 

Specific outcomes of the project include the following:

• Correlational analysis of participant experiences yielding overtime behavioral patterns affecting research results


Biology-Based Retinal Firing Algorithm for Model-Free Object Detection

• Counterfactual analysis of methodologies that are likely to prove effective
• Factorial design of Innovation House alternative formulations
• Multi-area lessons learned
• Two breakthrough technical capabilities

–Haptic data exploration
–Retinal firing algorithm for object discrimination

• Two new dot-coms
• Six research papers