Nuclear Nonproliferation in the Era of “Big Data”

Nuclear Policy Working Group (NPWG) Expert Panel:

April 7, 2015

Below are notes I took from talk through Nuclear Policy Working Group (NPWG) programming. The expert panel spoke about their expertise in nuclear nonproliferation efforts in the United States.

Panel Members:

Frank Pabian – LANL, satellite imagery, Nuclear Chief Inspector in Iraq for IAEA (1996-1998)

James Kornell – research engineer at National Security Technology’s Special Technologies Lab, computer ontologies, artificial intel, cognitive psychology.

Bryan Lee – director of Nonproliferation Studies, Army, Eurasian affairs

Maynard Holliday – assistant to undersecretary of Defense for AT&L (Pentagon), national security and acquisition policy, robotics, LLNL and Sandia Manager, start up robotics companies

Information available thru Internet:

  • passive collection vs active collection
  • crowd source
  • social media and other sources

Used tools (Sandia) – language indicators in programs

  • track shipments of nuclear materials

“layman” is the new possible expert, especially with open source info and data

easy to hide nuclear material (just some water!)

“red balloon hunt” – ex. of open source finding

“Boston bombing” – people find terrorists, but found wrong people

programming languages – free now

How much trust can you put into public as gatherers of info?

  • a lot
  • but data may not be judged by public for what it may be
  • medical indicators:
    • tweeting sickness – find secret chemical lab?
    • google earth – labeling sites
  • volume of pics taken is a lot! people take pics of “interesting” things – nuclear stuff may not be recognizable to general public

Social, Ethical considerations for using public data?

  • we are at beginning of how internet is impacting society
  • noncitizen info (ex. satellite data from all the time)
    • anonymous – nobody can be blamed in acquiring data/info.
  • NTI website on societal verification
    • other countries make it hard (treaties) to gather data from them

Lesser known sources for data nonproliferation:

  • wiki mapping, labeling for google earth
  • ecosec – meta search engine for social media site with map – look at “deadzone” at nuclear center for example
  • open-source software
  • google earth thermal data at nuclear sites
  • foreign social media sites (most nuclear-related stuff is there, not in US)

New approaches to analyzing that data?

  • comp. sci, stats, mathematics, etc.
  • “deep learning” linguistics – computational linguistics with grammar and jargon

Can nuclear info/data destabilize countries?

  • yes. it can be damaging

Can open-source play a role in Iran?

  • if they’re hiding something, open-source may be able to find something.

Does open source/crowd source require luck? or more systematic?

  • lots of sensing in phones – most of the time spent to clean up data rather than analysis. 
    • not luck , but how do we separate what’s relevant from irrelevant?
  • tips help

How do deal with disinformation?

  • false positives are generated, but typically people will reveal it is false (open-source)
  • not worried false positives will lead to major issues.

How separate signal from noise as analyst?

  • where does it come from? source? what is info? expert opinion? (Journalistic analysis before publish)
  • outside US may be different – harder to analyze
  • people are smart – “they can create noise to mask signal”

Develop open-source techniques for examining advanced fuel cycle facilities?

  • google earth already looks – for example
  • can you put that in public?
  • can’t tell what will happen since techniques develop over time

If US government takes up crowd-sourcing, how much will be given back to public? nonproprietary or classified?

  • things are already intrinsically open source
  • harder to “close the door” than “leave it open”
  • inside classified community – debates happen on what should be unclassified

Can you use similar techniques like epidemiologists or missing person searches?

  •  can use negative “lessons learned” info.
  • yes

Is crowd sources ok with a bunch of laymen or is it better for trained experts to go through data?

  • both is good.
  • teams (global, diverse) can search

How add new open sources as they develop? How set up gates for verifiability?

Privacy Concerns?

  • acknowledge what issues are – address in public sphere
  • colossal conundrum – need to craft policies to address privacy. there are still issues

How instigate open-source use by public in places?

  • ex-pats off mainland who report
  • issue with countries with one to two lines out to world internet
  • want to make it easy as possible for people to participate

How can public intentionally get involved?

  • follow people who do nuclear nonproliferation things online
  • twitter – nonproliferation – post it . share it.

Challenges to using open-source

  • what to look for? can be unclear what to look for in data
  • ethical, legal, societal implications of open-source are unknown. (not necessarily technical issues)
  • Detector enable law enforcement (DELI) – NYC police officers have radiation detection equipment
  • Can you extrapolate DELI for citizens?
    • apps to look for gamma signatures, for example
    • citizens as sensors

How does their work contribute to open-source verification?

  • Frank – textbook chapter, diversity of sources (satellite data)
  • Maynard – next generation technology for potential threats (apps and things to empower citizens around the world – passively contribute data)
  • Bryan – Jeffery Lewis military work/input, CNS host world only free and open trafficking database (radiological sources), text mining project – create better algorithm detection, Alibaba world market (Ebay) trading illegal or should be regulated items – trying to track that.
  • James – detect chemicals, small sensing, build app to find important information (open source, published)

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: