DYNAMIC PAGE -- HIGHEST POSSIBLE CLASSIFICATION IS
TOP SECRET // SI / TK // REL TO USA AUS CAN GBR NZL
(S//SI) Dealing With a 'Tsunami' of Intercept
FROM:
,
Human Language Technology (S23)
Run Date: 08/29/2006
and Anita H. Kulman
(S//SI) Everyone knows that analysts have been drowning in a
tsunami of intercept whose volume, velocity and variety can be
overwhelming. But the Human Language Technology Program
Management Office (HLT PMO) can predict that in the very near
future the speed and volume of SIGINT will increase even
more, almost beyond imagination. And we are working on ways
to help analysts deal with it all.
(S//SI) Of the HLT PMO's five Strategic Thrusts, the one that
addresses this problem is High Speed/ High Volume. It must
deal with today's collection and must plan for tomorrow's. The
current collection environment is characterized by huge amounts of
data, coupled with severely limited capability to send material
forward, and extremely limited number of queries that exactly
describe messages of value. That means we are capable of finding
huge amounts of data, much of which is not what we really want,
and that we cannot send it all back for analyst processing.
(TS//SI) To plan for tomorrow, High Speed/ High Volume is in line
with changes in the overall NSA/CSS systems, particularly
TURBULENCE and TURMOIL because when they become a reality in
the near future, we can expect collection capabilities to increase
significantly. TURBULENCE is an umbrella cover term describing
the next generation mission environment that will create a unified
system. TURMOIL is a passive filtering and collection effort on
high-speed networks. This is designed to be flexible and can be
modified quickly to deliver data in analyst-ready form.
(S//SI) One of High Speed/ High Volume's first efforts is in
developing and implementing ways to push HLT capabilities
very close to the collection points of the SIGINT system. In
particular, HLT is about to demonstrate an operational prototype of
language identification for Special Source Operations (SSO)
Counterterrorism text targets running at line speeds (STM-16) at
the packet-level. Resources permitting, HLT analytic processors will
automatically generate content-based events for TURMOIL based
on language.
(S//SI) HLT processors will demonstrate the ability to characterize
very high speed channels based on content, thus enabling analysts
to task the SIGINT system to send back messages based on
information found in message content, not just on externals.
(Externals can be Signal Related Information (SRI) that comes with
each message, such as channel, Time Up/Time Down, etc.) Using
HLT services, analysts will be able to build more precise
descriptions of the data they want. In addition, content-based
metadata will allow SIGDEV analysts to run more detailed surveys.
HLT services that work on data content at the collection point can
also provide indications or warnings that the SIGINT system must
adapt its collection strategy.
SERIES:
(U) HLT
1. Human-Language
Technology in Your
Future
2. For Media Mining,
the Future Is Now!
3. For Media Mining,
the Future Is Now!
(conclusion)
4. 'Knowledge
Discovery': Finding
the Best Material
5. Human-Language
Technology -Everywhere
6. Dealing With a
'Tsunami' of
Intercept
7. Building HumanLanguage
Technology
8. Strangers in a
Strange Land?
(S//SI//REL) Resources permitting, High Speed/ High Volume will
deploy capabilities for voice, text, and image data, and will
take advantage of research being done by a number of
organizations including the Research Directorate's Coping With
Information Overload Office (R6), Disruptive Technologies Office
(DTO), and SID/ Analysis and Production's Advanced Analysis
Laboratory (AAL). HLT research and transfer of its technology into
operations means the development of algorithms that can
incorporate HLT capabilities for the processing of elements such as
email attachments and VOIP.
(S//SI/REL) The research and technology transfer also may provide
"stealthy," low-profile in-target implants for Tailored Access
Operations (TAO) or technologies to enable high speed processing
in very low size, weight and power applications for other CLANSIG
customers. And, to help address the "unknown unknown" target
analysis problem, HLT is investigating techniques and technologies
for high volume voice processing so that all voice data can be
scanned for key words before it is selected based on phone
numbers.
(S//SI) Ultimately, HLT's High Speed/ High Volume will give the
analyst greater ability to influence collection and processing much
farther forward in the SIGINT system, as well as help the SIGINT
system achieve greater overall filtering and selection effectiveness.
That means more analysts wil be getting better SIGINT at a
time when volume and velocity are maximum.
"(U//FOUO) SIDtoday articles may not be republished or reposted outside NSANet
without the consent of S0121 (DL sid_comms)."
DYNAMIC PAGE -- HIGHEST POSSIBLE CLASSIFICATION IS
TOP SECRET // SI / TK // REL TO USA AUS CAN GBR NZL
DERIVED FROM: NSA/CSSM 1-52, DATED 08 JAN 2007 DECLASSIFY ON: 20320108