The Nine Dot Problem

1
Revelatory Attributes of Neuronal Group
Architecture for Strategic Knowledge
Management Harold E. Klein Dept. of
General Strategic Management Temple
University UCS 2007 May 14, 2006
2
Presentation Contents

• My focus Strategic Thinking The Situation
• An example of The Situation
• The Cognitive Problem
• Triggering a response in the brain
• How the brain handles complexity
• From a neurocognitive perspective
• Key ideas that can enhance strategic thinking
• Biological network architectures, various
• Optimal architecture for information organization
• Normative Cognitive Maps protocol and
  application, and
• NCM kinship with neuronal group architecture

3
My abiding interest Strategic thinking

• The Situation highly unstructured
• Multiple possible decisions or activities
  (possibly interdependent)
• Multiple factors, events, conditions (possibly
  causally or casually, directly or indirectly)
  affecting decisions/activities
• All, in a state of flux, over long time horizons

4
My abiding interest Strategic thinking

• The Situation highly unstructured
• Multiple possible decisions or activities
  (possibly interdependent)
• Multiple factors, events, conditions (possibly
  causally or casually, directly or indirectly)
  affecting decisions/activities
• All, in a state of flux, over long time horizons
• A really, really complex system!!!

5
Heres an example of the situation
6
From the NYT, April 28, 2004 Soon after 9/11, a
two-man intelligence team set up shop in a
windowless, cipher-proof room at the
Pentagon By the end of the year the men had
constructed a startling new picture of global
terrorism.
7
The men culled classified material We
discovered tons of raw intelligence So how were
these tons processed???
8
They recorded and annotated their evidence on
butcher paper hung like a mural around their
small office. (Accomplished over a period of
more than two months!!!) Is this any way to fight
the War on Terrorism???
9
Strategic thinking

• The Cognitive Problem -- How do you get people to
  do it when no one really knows how to do it
• -- or how it takes place (that is, in the mind
  of the decision maker)
• -- or what it is that triggers/stimulates a
  creative response
• The goal of a strategic planning process
  should not be to make strategy but to build
  prepared minds that are capable of making sound
  strategic decisions.
• (Sloan Management Review, Winter 2003)

10
Minds prepared with what?

• Deep seated biases/ingrained positions?
• Limited Experience?
• Executive hubris/self-confidence/self esteem?
• Knowledge homogeneity?
• Adverse disposition?
• Cultural baggage?

HOW IS SUCH PREPARATION OVERCOME?
11
Need to leap frog cognitive constraints-- so
that one can see the light
WITH GREAT DIFFICULTY!
12
To trigger/stimulate/create a strategic response

• It all stems from memory
• Learning is only a process of recollection
• Socrates in Platos Meno
• The synthesis of information/knowledge in the
  brain
• Information fusion

13
To trigger/stimulate/create a strategic response

• It all stems from memory
• Learning is only a process of recollection
• Socrates in Platos Meno
• The synthesis of information/knowledge in the
  brain
• Information fusion

14
To trigger/stimulate/create a strategic response

• It all stems from memory
• Learning is only a process of recollection
• Socrates in Platos Meno
• The synthesis of information/knowledge in the
  brain
• Information fusion
• Representation visualization key to what we
  see
• Evolved (or state) of information organization
  in nature

Discovery?
15
Where are the 123 separate HS-related agencies?
16

• And related and/or comparable situations at
• DHS responsive network organization
• CIA integrate information from different
  organizational units
• NSA any need to explain?

17
CIA Directorate of Intelligence
Regional offices 3 Transnational offices 6
Policy collection support 2
18

• So whats the solution?
• A representation of the situation --
• reflective of how the brain processes
  information
• decomplexifies
• revelatory
• In a manner that triggers/stimulates a creative
  response

19
How the brain handles complexity When confronted
by complex, nonroutine cognitive tasks --- the
brain is mainly a serial processor The more
complex the decision issue, the more factors to
be taken into account the greater the need
to map out a serial protocol with explicit
triggers to evoke the desired cognitive response
(i.e. extrinsic stimuli to memory)
20
From a Neurocognitive Perspective Butcher paper
murals set up an insoluble problem-solving
construct for the brain -- multiple factors to
be considered simultaneously -- leads to a
mental dead-end! The functional fixedness
phenomenon (Duncker, 1972) Unless problem is
reconfigured, it cant be solved (Reisberg,
1997) Efficacy of mental activity decreases when
confronted by two tasks (DEsposito et al., 1995,
Baddeley, 1997)
21
Efficacy of mental activity decreases when
confronted by two tasks (DEsposito et al., 1995,
Baddeley, 1997) Cognitive tasks below are
performed in serial fashion by the brain! Memory
recall (Sternberg, 1966) Object search
(Treisman et al., 1985) Processing simultaneous
stimuli (Kastner et al, 1998)
22
Memory the core of all thinking Data
storage in the brain is clustered, based on some
commonality.

• And there are other memory attributes
• Association
• Attention
• Disposition
• Perception
• Insight
• Imagery

Source Cognitive Memory Cellular and Network
Machineries and their Top Down Control, Science
10-24-04
23
Enhancing strategic thinking Role of mental
representation of problems extensively explored
(c.f., Cooper, 1980) The mental image evoked
by some analytical framework drives the mental
representation (or map) (Kosslyn, 1991) Choice
of a problem-solving protocol in the brain --
shown to take place from reconfigured
visualization of a structured problem
(Antonietti, 1991) The key then is the evocation
of the right mental imagery!!
24
Key ideas to consider---
We cannot understand a whole without seeing its
parts, but we can see the parts without
comprehending the whole. Thus, we may advance
from a knowledge of the parts to the
understanding of the whole.


Polanyi (1964) -Decision makers (anybody) have
difficulty simultaneously processing multiple
variables much less entire complex and
constantly adapting systems
25

• So, have humans contribute parts and use
  technology to assemble the whole
• But in a way that is revelatory!
• Visual representation of a complex whole is
  easier to comprehend than a verbal one.
• All visual representations are not created
  equal.

26
On comparing the character of the scenery in
different countries It depends chiefly on an
acquaintance with the individual parts of each
view. I believe that as in music, the person
who understands every note will more thoroughly
enjoy the whole, so he who examines each part of
a fine view, may also thoroughly comprehend the
full and combined effect. Charles Darwin
(1845) The trouble is the greater the number of
parts, the harder it is to comprehend the full
and combined effect (for most people).
27

• For any information system is to be used --
• - Most important of all, decision makers must
  own the product if it is to be seen as
  credible.
• Encourage a confrontation of individuals with
  unpleasant situations without allowing
  confrontation among individuals.
• - So, the protocol for evoking the situation
  representation is key to effectiveness

28
SOME ALTERNATIVE BIOLOGICAL NETWORKS
29
Typical artificial neural network
Source Spontaneous Evolution of Network Motifs,
PNAS 092705
30
Neuronal-type architecture -- not limited to the
brain
The network model shows how the various
modules interact with each other to coordinate
and regulate cellular activity.
Protein-Protein interaction patterns within a
yeast cell Source A data integration methodology
for systems biology, PNAS 11-29-05
31

• Complex Network Models schematics (left) 256
  node, scale-free networks (right)
• Characterized by
• A few highly connected nodes or hubs (blue dots)
• B) Four highly interlinked modules connected to
  each other by a few links
• Hierarchical topological organization with
  embedded modularity hierarchical levels are
  represented in increasing order from blue to
  green to red (note that the networks underlying
  modularity is hard to discern)
• Source Hierarchical Organization of Modularity
  in Metabolic Networks, Science 08-30-02

32
Network Analysis of Committees Subcommittees of
the 107th U.S. HR (2001-2002)
WHAT DOES THIS ANALYSIS TELL YOU? LINKS ARE
INTERLOCKS. SO WHATS BEING MEASURED?
33
Network Analysis of Committees Subcommittees of
the 107th U.S. HR (2001-2002)
Link between nodes measures the extent of common
membership in two committees Source Porter et
al., PNAS, 05/17/05
34
Brain activity pattern or map
Whats missing?
Source Synchronous dynamic brain networks
revealed through MEG PNAS 01-10-06
35
Brain activity pattern or map
Whats missing?
Functional orientation of individual local
regions
Source Synchronous dynamic brain networks
revealed through MEG PNAS 01-10-06
36
Clustering Analysis of Protein Interactive
Networks
Clusters comprised of proteins with physical /or
logical interactions to an identical set of
proteins, also enhanced with functional modules
Source B. Andeopoulos et al., Bioinformatics
23, 9 2007
37
MY EPIPHANY
38

• Edelmans proposed neuronal group architecture
• group structure triggered by common set of
  inputs
• all neuronal groups not connected to all inputs
• No two groups internally structured the same way
• More interconnectedness within groups than among
  groups
• Source Bright Air, Brilliant Fire On the Matter
  of the Mind

39
An optimal architecture of knowledge management?

• If a series of dynamic relationships are
  identified (in the general form of Node A
  influences Node B which, in turn, has relevance
  for Subject 1), then the following operations can
  be performed
• Cluster closely interacting nodes that have
  relevance for commonly affected
  subjects/issues/decisions
• Cluster subjects/issues/decisions that are
  commonly affected by the same nodes
• A node can be any defined thing.

40
The result when displayed visually

• An architecture that is remarkably similar to
  that of Edelman
• I call this Normative Cognitive Maps

41
NORMATIVE COGNITIVE MAPS (NCM) Originally
designed for L/R strategic planning for Exxon.
Next applied at IBM Celanese. Later at Douglas
A/C Florida Power Light
42

• WHAT IS NORMATIVE COGNITIVE MAPPING?
• A strategic decision support system for large,
  complex organizations, characterized by
• - Multiple strategic decision issues (SDIs)
• - Complex interdependencies among SDIs
• - Extensive coordination required among SDIs
  and decision makers
• - Shifting, unstable, chaotic environments
• A collaborative technology
• - Allowing mutually exclusive inputs from any
  number of sources

43

• THE NCM APPROACH
• A pc-based protocol for representing the
  organization's relevant environment in a unique
  causal mapping format of immediate use to
  strategic decision-makers.
• NCM inputs are individual statements of
  hypotheses, conjectures, forecasts, analyses
  concerning the prospective organization/environmen
  t interface
• NCM utility is best shown by example

44
FPL APPLICATION - MOST EXTENSIVE TO DATE
(EPRI-SPONSORED)

• FPL selected for NCM demonstration project
  because of companys bellwether status
• as leading edge corporation in utility industry
  management style, and
• in the midst of reinventing itself

45

• FPLs environmental reassessment
• Under guidance of outside consultant
• Organized 12 cross functional, cross hierarchical
  committees (10-20 per) working for six months
• Each committee focusing on a major SDI (e.g.,
  generating capacity, fuel supply, transmission,
  regulation), producing forecasts, scenarios, etc.
• Consultant distilled and synthesized committee
  results into several hundred pages of cross
  referenced text, tables and graphs

46

• NCM application at FPL
• FPL committee data summaries were inputs to the
  NCM approach
• Distilled down to 52 verbal statements of
  environmental dynamics with relevance for FPL
  SDIs
• Consisting of 66 specific environmental factors
  and 34 SDIs
• Verbal statements transformed into NCM input
  protocol (environmental dynamics/SDI linkage
  statements)

47
(No Transcript)
48

• Run the NCM heuristic program w/notational
  statements input
• Identifies and groups closely interacting
  environmental dynamics
• With relevance for commonly affected SDIs
• Five subsets of environmental dynamics/SDIs
  emerged
• Each subset provides the specification for a NCM
  diagram

49
(No Transcript)
50
(No Transcript)
51
(No Transcript)
52
(No Transcript)
53
Schematic of NCM Diagrams
54

• Edelmans proposed neuronal group architecture
• group structure triggered by common set of
  inputs
• all neuronal groups not connected to all inputs
• No two groups internally structured the same way
• More interconnectedness within groups than among
  groups

55

• NCM diagram architecture resembles biological
  neural networks, exhibiting
• plasticity
• connectivity
• emergence
• self-organization

56

• NCMs emulated neuronal networks both in
  behavior and architecture
• Stimulus-response linkage
• Unidirectionality
• Pairwise connections
• Clustering
• Plasticity
• Connectivity, and
• Internal neuronal group network
• structure

57
Thanks for your attention! Questions? Comments?
58

• Edelmans proposed neuronal group architecture
• group structure triggered by common set of
  inputs
• all neuronal groups not connected to all inputs
• No two groups internally structured the same way
• More interconnectedness within groups than among
  groups

59
WHAT NCM PROVIDES

• ANSWERS TO KEY QUESTIONS ASKED OF PLANNERS
• What strategic decisions need to be coordinated
  how do we organize to do so
• What is the sequence or order in which decisions
  need to be addressed
• Where can we intervene most effectively to shape
  our environment

60
In addition, NCM

• Provides a comprehensive tool for monitoring
  environmental change
• Produces sequence of event (or path) forecast
  scenarios with direct relevance for SDIs

61
Evolution of a neuron
62
(No Transcript)