DARPA Seeks to Unlock Big Data Mechanisms

The world is awash in data, and yet the ability to connect it all in meaningful ways seems to be drifting further out of reach. Enter the Defense Advanced Research Projects Agency (DARPA), which last month announced a new project aimed at building new models that drive insight into the inner workings of complicated systems, such as ecosystems, brains, and economic and social systems.

As part of its newly announced Big Mechanism project, DARPA will provide $45 million in grants to 12 award winners at universities, government, and private companies over a period of several years. The goal of the program is to develop technologies for “a new kind of science” in which research is integrated more or less immediately into “causal, explanatory models of unprecedented completeness and consistency,” the agency says. While the immediate goal is to find ways to detect signal pathways for cancer cells, the program has far wider implications.

“Big Mechanisms are causal, explanatory models of complicated systems in which interactions have important causal effects,” the agency says in its official 48-page Big Mechanism proposal. “The collection of Big Data is increasingly automated, but the creation of Big Mechanisms remains a human endeavor made increasingly difficult by the fragmentation and distribution of knowledge. To the extent that we can automate the construction of Big Mechanisms, we can change how science is done.”

Those chosen to work on DARPA’s Big Mechanism project will have overlapping disciplines, including natural language processing; curation and ontology; systems and mathematical biology; and representation and reasoning. The agency isn’t quite sure where all this will lead, but it expects skills in visualization, simulation, and “statistical foundations of very large causal networks,” may also come into play.

DARPA envisions advances being made in various parts of the sciences, not just oncology and bio-informatics. “Machine reading researchers will need to develop deeper semantics to represent the causal and often kinetic models described in research papers,” the agency says. “Deductive inference and qualitative simulation will probably not be sufficient to model the complicated dynamics of signaling pathways and will need to be augmented or replaced by probabilistic and quantitative models. Classification and prediction will continue to be important, but causal explanation is primary.”

In the end, the agency hopes to have a collection of tools that can:

1. Read abstracts and papers to extract fragments of causal mechanism;
2. Assemble fragments into more complete Big Mechanisms; and
3. Explain and reason with Big Mechanisms.

 If it all sounds a bit over-the-top, gonna-change-the-world, Tony Stark-ish to you, then you don’t know DARPA. The agency that actually created the Internet is setting out to change how data science is done.

Related Items:

Stanford Receives DARPA Grant to Study Big Data

Georgia Tech Scores $2.7 Million to Move DARPA Big Data Goals

How DARPA Does Big Data