Intelligent Search

Diffeo uses machine learning to streamline research and discovery


Software startup Diffeo (Booth 1623) wants to revolutionize the way we search for files and information by answering for its users the question: “How do I know what I missed?”

Diffeo’s cloud search and discovery capabilities help analysts discover “unknown unknowns.”

“How do you discover something you didn’t know you were looking for? I’m looking for information on [a topic]—I don’t know what specifically, but I want to know as much as I can. Diffeo shows those non-obvious relationships,” said Geoffery Milstein, VP of Federal Programs at Diffeo.

Diffeo grew out of the Text Retrieval Conference Knowledge Base Acceleration project, which it organized for the National Institute of Standards and Technology and DARPA Memex. The company also won the grand prize last year in the National Geospatial-Intelligence Agency’s Disparate Data Challenge.

At GEOINT 2017, the company is demonstrating how it uses “collaborative machine intelligence,”—or AI—in algorithms to help accelerate the discovery of pertinent information during research.

What differentiates Diffeo is that its software operates seamlessly within a user’s normal work process, namely in applications like Microsoft Word or OneNote. As progress is made and data is added, the machine performs textual analysis to continuously learn about the user’s interests and what sorts of disparate information might be useful. The software then sends automatic queries to external and internal sources like databases, SharePoint repositories, and the internet, pulling together relevant information for the user.

“You don’t have to have ground-truth data. Every time the user interacts with the machine, it is reformulating its model,” Milstein said. “It’s not just another tool you have to learn.”

Additionally, Diffeo is showcasing its “knowledge boards” or graph displays, which perform gap analysis between working documents and, for example, a website the user has been frequenting in order to create visual representations of the data.

Image courtesy of Diffeo


DNI Haines: GEOINT ‘Fundamental’ to U.S. National Security

Director of National Intelligence Avril Haines discusses the improvement and expansion of GEOINT—and where we go from here


DoD CIO Outlines Plans to Safeguard Data Amidst Conflict

CIO John Sherman points to zero trust, new cloud partnership as keys to operational success


A Reflection on 40 Years of Service

Lieutenant General Scott Berrier, director of the Defense Intelligence Agency, reflects on what has changed and what remains consistent in the threat landscape