Ms. Frances Fragos Townsend

Ms. Frances Fragos Townsend gave a good speech about supporting Homeland Security with Open Source Intelligence.

Technology: Improving the Use of Open Source

Agenda

Dr. Joe Markowitz, is talking about turning the analysis process upside down. He’s responsible for desktop secure KVM between multiple security domains. He’s original plan was to have people work at the lowest level. Gather the most information about topic X or target X on the lowest domain. Then move up….not the other way around. Open Source Intelligence should be the main go to source…first, not last, and should be an integrated part of the analysis process.

He’s also talking about cross domain information sharing…one way, up. Because if the analysis process is turned upside down, one way, up, is the easy way to go. Down domain transfer are the hard part. So he’s for working at the lowest level.

On audience member asked about website redirects based on IP address, whereas a USA internet user would be directed to version B of Hizballah website, and what is the Open Source collection analyst doing to avoid this. My first thought was tor.eff.org, the panelist gave a round about answer.

Another question about multiple logins, where as the user as valid access to multiple systems on the same domain. Where are we with integrating this. My first thought was the IC-PKI effort and Scattered Castles, whereas the PKI validates who you are and what access levels you have. The panelist only said there are pilots in place, and said [paraphrase] “you have different keys for your house, car, etc” I think he didn’t understood the question fully, and was thinking more about the importance of compartmenting information.

Lunch

Lunch for most was a ham sandwich, basically, but there are allot of people here.

Data Fusion and Data Enrichment Discussion

Mr. Voultepsis: Five themes: self services, timely, agile, robust and simple.

All discussions are paraphrased.

Q: What needs to be done to make machines read text better and what questions do we need to ask machines?

A: Dr. Popp: Clustering, Filtering and semantic abilities. Perhaps others. Technology to mimic a human brain.

A: Ms. Goldbeck: Don’t believe that we can automatically derive the same human understanding automatically. Humans will have to be involved.

A: Dr. Popp: NSA was producing so much info per day that only 1% was able to be processed by humans. How to get all this information looked at. How to get the 80% or 50% solution.

Q: How to ask the right questions?

A: Dr. Popp: Asking the right question is sometimes harder than doing the modeling work.

A: (from the back) Search strings, boolean logic, advanced technology to get the answer/

Q: How to integrate several closed data repositories?

A: Ms. Goldbeck: This is the basis of the Semantic Wen Technologies. Taking info from different databases and combine if you have basic ontologies.

A: Mr. Ryan: Multi-search across 13 different agency databases that are not structured alike.

A: Dr. Popp: The bigger challenge is not the etchnical one, but more the policy one, involving ORCON, and other classifications.

A: Mr. Voultepsis: One needs to be aware when there is more information out there.

Q: How quickly can you create a variation of ontology

A: Ms. Goldbeck: Very quickly. I can create one in an hour. Expansions are trivially easy. Large agencies spend more time and are still able to do monthly revisions. Smaller changes can be made within an hour or two on smaller data sets.

A: Dr. Popp: Agrees that ontologi es tend to be rigid and need to be able to adapt due to change.

End of session.

Data Fusion and Data Enrichment Continued…

Next up is Jennifer Goldbeck From the University of Maryland. She speaks on the Semantic Web, Provenance and Social Networks.

Basics for building an Open Source Portal on the Web:

1. Access info from all over the Web (text, video, photos, note sources)
2. Intelligence Analysis (conclusions added to portal, whose analysis, when, what used)
3. Filter information to based on trust and confidence

The actual technologies must include integration of distributed information. The other technology needs to track the provenance of the information.

Semantic Web: need technology to bring this machine processable information into a system.

Provenance tracking: the history and background information. The ability to gauge confidence of the information.

Social Networks: Who? How much do we trust that person? If we don’t know a person directly, can we compute how much to trust them?

Integration of Data, Provenance and Social Networking to help filter the information. The question is how to put this all together to make the most efficient work. http://profilesinterror.mindswap.org is an example from the U Maryland team working on this problem.

Next up is Mr. Ryan from DTIC (http://www.dtic.mil). DTIC has worked with Open Source and the Intelligence Community for a long time (DefensLink as an example). Mr. Ryan details DTIC’s history and current standing and holdings.

Drop-outs

The networks is dropping out at very inconvenient times (when I am trying to save a draft). Argh! I have had to reconnect three times in this discussion so far!

Data Fusion and Data Enrichment

The panel consists of Alex Voultepsis, Chief of Enterprise Services Division of the DNI , Dr. Robert Popp, CEO of National Security Innovations, Raul Valdez-Perez, CEO of Vivismo, Inc., Jennifer Golbeck, Asst Dir for the Center for Foreign Information Policy and Electronic Govt at U Maryland, and R. Paul Ryan, Administrator, Defense Technical Center.

Dr. Popp: Key problems are failed states, WMD and terrorism.

Types of technologies: Foreign Languages Tools, Analysis Tools and Decision Aides, Predictive modeling, Pattern Analysis, and Privacy Protection Technology. (lots of slides.)

Three fundamental challenges:

Can the analytic community imagine the many types of signatures that terrorist plans, plots and activities will create?

Can we detect these signatures when embedded in background noise?

Can we detect these signatures without adversely violating the privacy of non-terrorists?

Inverting the Analyst bathtub curve: What an analyst needs to do: research, analysis, production. Generally more time is spent on research and production of reports when it should be most time on analysis.

Among the examples shown, Dr. Popp shows how some predictive modelling tools can save money and man years.

Dr. Popp makes the point that R&D inData Fusion and Data Enrichment should not be curtailed. Oversight and checks and balances should be put into place but banning the technology is detrimental to analysis.

Next is Mr. Valdez-Perez of Vivismo. Some of you may be familar with Vivismo as the only clustering engine on the high-side as well as the search engine that is used with MDDS to search multiple networks at once. He discusses the overload of information and discusses clustering. He illustrates the clustering using Clusty, one of their products.

Open Source on the Web

The Polaris room, full of young to old individuals, a panel of gentlemen stand in the front next to a blue projection screen.

Agenda

Jeff Baxter and others talk about various Open Source topics. (very descriptive)

Questions for Sen Kerrey

Again, paraphrasing.

Q: Are there any models for the large decentralized repository you spoke of?
A: Community College based, media literacy is important, as long as it was open, to see who was doing good work, allow the community to decide who should get the funding. Organic funding. You need to put enough money in order for people to pay attention to it. Democracize it beyond the big guts (no offense to MIT and others) but that it needs to allow for more to be involved.

Q: What is your opinion on engaging the private sector in intelligence problems?

A: As long as the intelligence is defined. We are way short of providing (esp local levels) is training and that we are short in building the interoperable network that we need to get to. Health, communication, transportation.

Q: What are some successes, besides the conference and the ADDNI/OS office?

A: The quality of people involved since 9/11. Volunteers to the cause in the past 6 or 7 years. Pockets of excellence in various areas of the intelligence community. Not pessimistic, but the debates tend to be less informative than they need to be but that they get wearying. We as citizens are too impatient and we need to give the young people the authority to tell us what to do. The agencies can’t be the same as in 1967. He has no difficulties finding areas that cause him to be optimistic.

Q: How do we get around the beauracracy and contract process?

A: There is no “perfect”. Old national security acts (1947) do not apply to the digital world. We need to come into it with the right  sense of urgency. Sec Rumsfeld’s Sept 10, 2001 speech addresses this.

Q: How do we get the State and Local entities engaged in the national problem.

A: We are seeing a lot of attention paid at the municipal level. This needs to be addressed at the National Governor’s Association. They are already working on a lot of problems and this is not going to come to the forefront.

More from Former Senator Kerrey

He discusses the computer and technical talent of teenagers and youth – in the Islamic world.

He mentions the importations of food products (as well as pet food and toothpaste) from China. China recently executed a corrupt official responsible for sub-par foods and medicines being allowed on the market, but it doesn’t allay the worries Americans have over the recents problems with these imports.

He discusses Congress as the ultimate customer of Open Source so that they understand before making decisions. He mentions how there are too many possibilities and expertises that members of Congress cannot possibly be experts on all of them. He says that we need to deliver to Congress the type of tool to allow understanding of groups, issues and networks. He would like to see some type of network that makes it possible to decentralize information. He would like to see a research project on this. And he says that it is absolutely necessary to internationalize the process. Over there is here.  And the fear of new knowledge needs to be conquered. And we need to tap into the research of the social scientists so that we can understand more and that it isn’t put on the shoulders of a few analysts.