Cognitive Triangle. Dec The Overall classification of this Briefing is UNCLASSIFIED

Transcription

1 Cognitive Triangle Dec 2012 THIS INFORMATION IS FURNISHED WITH THE UNDERSTANDING THAT IT IS TO BE USED FOR DEFENSE PURPOSES ONLY; THAT IT IS TO BE AFFORDED ESSENTIALLY THE SAME DEGREE OF SECURITY PROTECTION AS SUCH INFORMATION IS AFFORDED BY THE UNITED STATES; THAT IT IS NOT TO BE REVEALED TO ANOTHER COUNTRY OR INTERNATIONAL ORGANIZATION WITHOUT THE WRITTEN CONSENT OF THE UNITED STATES GOVERNMENT DEPARTMENT OR AGENCY FURNISHING THE INFORMATION. The Overall classification of this Briefing is UNCLASSIFIED

2 Name of the Game Too much information Too fast for humans not enough humans Competitive environment Computers to the rescue! Information management Collect & process information, all that you can Transform into decision & action The complex (static and dynamic), competitive environments capitalism, war, sports, casinos, disaster relief Increasingly both are being dominated by autonomous systems 2

3 By Chris Steiner Flash crash of 6 May 2010 Oakland A s story in Moneyball Computer-Assisted Legal research (CALR) NYT Science Times: December 4, 2012 For Second Opinion, Consult a Computer? 3

4 Example Algorithmic trading is the use of an algorithm for entering trading orders and deciding on aspects of the order such as the timing, price, or quantity of the order, or in many cases initiating the order without human intervention. A special class of algorithmic trading is high frequency trading (HFT), in which computers make elaborate decisions to initiate orders based on information that is received electronically, before human traders are capable of processing the information they observe. 4

5 Complex event processing Complex event processing, or CEP, is event processing that combines data from multiple sources to infer events or patterns that suggest more complicated circumstances. The goal of CEP is to identify meaningful events (such as opportunities or threats) and/or patterns and respond to them as quickly as possible CEP is used with business process management (BPM) to focus on end-to-end business processes, in order to continuously optimize and align for its operational environment. 5

6 CEP Solutions Most CEP solutions and concepts can be classified into two main categories: Computation-oriented CEP Executing on-line algorithms as a response to event data entering the system. Detection-oriented CEP detecting combinations of events called events patterns or situations Creates actionable, situational knowledge but not decisions or actions Uses the computer and the network part of the triangle A more complete process is the OODA Loop 6

7 OODA The OODA loop (for observe, orient, decide, and act) is a concept originally applied to the military operations and is now also often applied to commercial operations and learning processes 7

8 OODA Loop Operations Designed to describe a competition The key is to obscure your intentions and make them unpredictable to your opponent while you simultaneously clarify his intentions. Operate at a faster tempo to generate rapidly changing conditions that inhibit your opponent from adapting or reacting to those changes and that suppress or destroy his awareness. 8

9 The Cognitive Triangle Natural Language Dialogue System Sensing/Perception Machine Learning Cognitive & Social Science Issues Software Agents Multi-Agent Systems Machine Brain-Computer Interface

10 Goal How does one engineer the cognitive processes and distribution to maximize overall system metrics such as speed of data input-to decision-output or reduction of errors and false alarms? What has cognitive science/engineering discovered concerning interacting with computer-based information systems? Are there cognitive processes in the human brain that can be emulated or simulated (by computer) which, when combined in the triangle, will improve throughput, bandwidth or some other system metric? 10

11 Cognitive Triangle and the OODA Loop Machine Machine Machine Machine 11

12 Possible Scenario 1 5 Day Disaster Relief Race against time Natural disaster (earthquake/tsunami) hits populated regions around part of the Pacific rim. Humanitarian relief must be provided ASAP including: Survivor search Hospital services Food, water, shelter Basic energy infrastructure (to run hospitals, transportation) ASAP is to prevent disease (it s a race)---5 days 12

13 Possible Scenario 2 Competition Football rivals: World Cup, USA vs Brazil Factors to be considered: Team roster Training History of performance, individual and group Tactics, techniques, and procedures Goal is to assemble and run a team throughout the season designed to win the world cup against Brazil 13

14 1. Cognitive Load Distribution Function Machine Network Human Observe X%, X Y%, Y Z%, Z Orient Decide Act For each OODA loop function.how is the cognitive load distributed? What enabling tech must be improved to optimize the distribution? How and by what delta % wil the load change due to expected emerging tech improvements? 14

15 Enabling Technologies First consideration is to systems-integration challenges: Don t stovepipe the enablers Mastery of individual enablers does not ensure success, integration of enablers does Determine key research goals in the enabling technology areas as they relate to OODA Loop performance. List the relevant milestones that must be reached enabling technologies List the relevant impediments to achieving technological breakthroughs in these areas 15

16 2. Performance Improvement The ultimate metric of interest is the time through the OODA loop What is the best performance improvement (time) that can be expected given anticipated improvements in enabling technologies? Substitute is % improvement based on enabling tech improvement 16

17 Tasks Select a scenario or scenarios Determine current performance (sliding scale?) Optional in the sense that deltas are acceptable Examine the enabling technologies to determine load and % improvement in the context of OODA Loop application No stovepipes Determine the performance metric and deltas 17

18 Bottom Line To transit the OODA loop: What improvements must be made in each enabling tech area? By how much? How much performance improvement can be expected in the OODA loop? 18