1. Cognitive Architecture
B. Kiran Maruthi 09005047
M. Sonika 09005054
D. V. Ramana 09005059

2. OUTLINE
 What is Cognitive Architecture?
 Plausibility of Cognitive Architectures
 Type­Identity Theory
 Functionalism
 History of Cognitive Architecture
 General Characteristics
 Consciousness
 Unified Theory of Cognition
 SOAR – case study

3. Intelligent Agents
 Entities which observe through sensors and act upon the
environment using actuators and direct their activity towards
achieving goals.

4. What is Cognitive Architecture?
 Blueprint for intelligent agents.
 It proposes (artificial) computational processes that act
like cognitive systems (human)
 An approach that attempts to model behavioral as well as
structural properties of the modeled system.
 Aim : to model systems that accounts for the whole of
cognition, i.e., systems with Artificial Consciousness –
which can not only respond but also think, perceive and
believe like a human !

5. Artificial Consciousness
 Artificial Consciousness is broadly classified
as access and phenomenal consciousness.
 Brain processes neural impulses from
the eyes and determines that this image is
physically unstable – pattern recognizability.
 What about pain, anger, motivation, attention, feeling of
relevance, modeling other people's intentions, anticipating
consequences of alternative actions, or inventing ?

6. Plausibility of Artificial
Consciousness
 A view skeptical of AC is held by type­identity theorists
“consciousness can only be realized in particular
physical systems because consciousness has properties that
necessarily depend on physical constitution”
 However, for functionalists,
“any system that can instantiate the same pattern of
causal roles, regardless of physical constitution, will
instantiate the same mental states, including
consciousness”
Along these lines, some theorists have proposed that
consciousness can be realized in properly designed and
programmed computers.

7. Type­Identity Theory
 The mental events can be grouped into types and
associated with types of physical events in the brain.
 For example, mental event pain results in physical event in
the brain (like C­fiber firings)
 We have two totally different versions of type­identity
theory based on the definition of “what kind of identity” is
associated with mental and physical events.
­ Ullin Place (1956) – Compositional Identity
­ Feigl (1957) and Smart (1959) – Referential Identity

8. Compositional Type­Identity Theory
 U.T.Place's notion of identity is described as a relation of
composition.
 Every mental process is composed of a set of physical
sensations to which it reacts. But can we associate them
based purely on composition?
 "lightning is an electrical discharge" is true.

9. Referential Type­Identity Theory
 For Feigl and Smart, the identity was to be interpreted as
the identity between the referents of two descriptions
which referred to the same thing.
 “the morning star” and “the evening star” are identical in
the sense that both of them refer to the Venus.
 Sensations and brain processes do indeed mean different
things but they refer to the same physical phenomenon.
This is called as The Fregean distinction
 Conclusion : All of the versions share the central idea that
the mind is identical to something physical.

10. IMPOSSIBRU!!!

11. Multiple Realizability
 Objections to the type
identity theory
 Hilary Putnam popularized it
in late 1960s.
 It states that the same
mental property, state, or
event can be implemented by
different physical properties,
states or events.

12. Putnam's Formulation
 Do all organisms have the same brain structures? Clearly
not !
 Pain corresponds to completely different physical states and
yet they all experience the same mental state of "being in
pain."
 Should robots be considered a priori incable of expereincing
pain just because they did not posses the same
neurochemistry as humans?
 Putnam concluded that type­identity is making an
implausible conjecture.

13. Functionalism
 Core idea is that mental states are constituted solely by
their functional role
 They are causal relations to other mental states, sensory
inputs, and behavioral outputs.
 Brains are physical devices with neural substrate that
perform computations on inputs which produce behaviours.
 According to this theory it is possible to build silicon based
devices which are functionally isomorphic to the
humans as long as system performs appropriate functions.

14. Variations of Functionalism
 Machine State functionalism – Hilary Putnam
 Mental state is like automaton state of a Turing Machine.
 Each state can be defined exclusively in terms of its
relations to the other states as well as inputs and outputs.
 Being in pain is the state which disposes one to cry "ouch"!

15. Variations of Functionalism­Cont...
 Psycho functionalism – Jerry Fordor
 The role of mental states, such as belief and desire, is
determined by the functional or causal role that is
designated for them within our best scientific psychological
theory.
 If some new mental state from folk psychology comes,it is
considered non­existent as it has no fundamental role in
cognitive psychological explaination.
 Some theoretical cognitive psychological states which are
necessary for explaination of human behaviour but are not

16. Quick Question
What difference does the colour RED make?

17. Qualia
 From the Latin, meaning "what kind".
 refers to the subjective qualities of sensory perception and
the feeling they generate.
 Qualia is not only the “redness” of red, but the way that
redness makes us feel.
 Qualia are, in essence, our own unique and personal
perceptions of our environment.

18. Mary's thought experiment
 Frank Jackson offers the knowledge argument for qualia.
 Mary, the colour scientist knows all the physical facts
about colour and the experience of colour with other people.
 Confined from birth to a room that is black and white.
 When she is allowed to leave the room, it must be admitted
that she learns something about the colour red the first
time she sees it — specifically, she learns “what it is like”
to see that colour.
 This attacks the knowledge completeness of functionalism.

19. Ability Hypothesis
 Nemirow claims that "knowing what an experience is like
is the same as knowing how to imagine having the
experience".
 He argues that Mary only obtained the ability to do
something, not the knowledge of something new.
 Mary gained an ability to "remember, imagine and
recognize."
 Knowing what it's like to see red is merely a sort of
practical knowledge, a “knowing how” (to imagine,
remember, or re­identify, a certain type of experience)

20. Functional Isomorphism
Putnam defined the concept of functional isomorphism as :
Two systems are functionally isomorphic if there is a
correspondence between the states of one and the states of
the other that preserves functional relations.

21. Presently...
Functionalism is widely accepted and research to develop
cognitive robots is on!

22. Cognitive Architecture
 Using Putnam's Multiple Realizability formulation and
functionalism, David Chalmers in late 1960s suggested the
possibilty of mechanisms and structures that underlie
Cognition :
­ processors that manipulate data
­ memories that hold knowledge and
­ interfaces that interact with an environment.

23. History of Cognitive Architecture
1969­2000(time line)

24. • GPS (Ernst & Newell, 1969) Means-ends analysis, recursive subgoals
1970 • ACT (Anderson, 1976) Human semantic memory
• CAPS (Thibadeau, Just, Carpenter) Production system for modeling reading
1975 • Soar (Laird, & Newell, 1983) Multi-method problem solving, production systems, and problem spaces
• Theo (Mitchell et al., 1985) Frames, backward chaining, and EBL
1980 • PRS (Georgeff & Lansky, 1986) Procedural reasoning & problem solving
• BB1/AIS (Hayes-Roth & Hewitt 1988) Blackboard architecture, meta-level control
1985 • Prodigy (Minton et al., 1989) Means-ends analysis, planning and EBL
• MAX (Kuokka, 1991) Meta-level reasoning for planning and learning
1990 • Icarus (Langley, McKusick, & Allen,1991) Concept learning, planning, and learning
• 3T (Gat, 1991) Integrated reactivity, deliberation, and planning
1995 • CIRCA (Musliner, Durfee, & Shin, 1993) Real-time performance integrated with planning
• AIS (Hayes-Roth 1995) Blackboard architecture, dynamic environment
2000 • EPIC (Kieras & Meyer, 1997) Models of human perception, action, and reasoning
• APEX (Freed et al., 1998) Model humans to support human computer designs

25. Characteristics
 Holism, e.g. Unified theory of cognition
 The architecture often tries to reproduce the behavior of
the modelled system (human), in a way that timely
behavior (reaction times) of both are comparable
 Other cognitive limitations are often modeled as well
 Robust behavior
 Parameter – free
 Artificially Conscious

26. Artificial Consciousness
The functions of consciousness suggested by Bernard Baars :
 Definition and Context Setting
 Adaptation and Learning
 Anticipation Function
 Prioritizing and Access­Control
 Decision­making or Executive Function
 Analogy­forming Function
 Metacognitive and Self­monitoring Function
 Autoprogramming and Self­maintenance Function
 Definitional and Context­setting Function.

27. Learning
 Reaction time for consecutive readings?
 Human improvement via Practise

28. Anticipation
 Machine needs flexible, real­time components that predict
worlds.
 A conscious machine should make coherent predictions and
plans, for environments that may change.
 Executed only when appropriate to simulate and control the
real world.
 Significant research on role of consciousness in cognitive
models. Examples : CLARION, OpenCog

29. Unified Theory of Cognition
 Book written by Allen Newell
 Newell's goal :
To define the architecture of human cognition, which is the
way that humans process information. This architecture
must explain how we react to stimuli, exhibit goal directed
behavior,acquire rational goals, represent knowledge, and
learn.

30. Newell's Cognitive Model
 Newell introduces Soar, an architecture for general
cognition.
 Soar is the first problem solver to create its own subgoals
and learn continuously from its own experience.
 Soar has the ability to operate within the real­time
constraints of intelligent behavior, such as immediate­
response and item­recognition tasks.

31. Soar
 What is Soar?
 History of Soar
 Architecture of Soar
 Evolution of Soar and present version

32. What is Soar?
 Soar is a symbolic cognitive architecture.
 An AI programming language.
 It provides a (cognitive) architectural framework, within
which you can construct cognitive models.
 It can be considered as an integrated architecture for
knowledge­based problem solving, learning, and interaction
with external environments.

33. History
 Created by John Laird, Allen Newell, and Paul Rosenbloom
at Carnegie Mellon University in 1983.
John Laird Allen Newell Paul Rosenbloom

34. It's Soar not SOAR !
 Historically, Soar stood for State, Operator And Result,
because all problem solving in Soar is regarded as a search
through a problem space in which you apply an operator to
a state to get a result.
 Over time, the community no longer regarded Soar as an
acronym: this is why it is no longer written in upper case

35. Screenshot – Soar Debugger

36. Problem Spaces
 Soar represents all tasks as collections of problem spaces.
 Problem spaces are made up of a set of states and operators
that manipulate the states.
 Soar begins work on a task by choosing a problem space,
then an initial state in the space. Soar represents the goal of
the task as some final state in the problem space.

37. Structure of Soar
 Soar can be divided into 3 levels :
 Memory Level
 Decision Level
 Goal Level

38. Memory Level
 A general intelligence requires a memory with a large
capacity for the storage of knowledge.
 A variety of types of knowledge must be stored, including :
­ declarative knowledge
­ procedural knowledge
­ episodic knowledge

39. Long­term Production Memory
 All of Soar's long­term knowledge is stored in a single
production memory.
 Each production is a condition­action structure that
performs its actions when its conditions are met.
 Memory access consists of the execution of these
productions.
 During the execution of a production, variables in its
actions are instantiated with value.

40. Working Memory
 The result of memory access is the retrieval of information
into a global working memory.
 It is the temporary memory that contains all of Soar's
short­term processing context. It has 3 components :
­ The context stack specifies the hierarchy of active goals,
problem spaces, states and operators
­ objects, such as goals and states (and their subobjects)
­ preferences that encode the procedural search­control
knowledge

41. Soar
Architecture

42. Preferences
 There is one special type of working memory structure ­
“the preference”
 Preferences encode control knowledge about the
acceptability and desirability of actions.
 Acceptability preferences determine which actions should
be considered as candidates.
 Desirability preferences define a partial ordering on the
candidate actions.

43. Decision Level
 The decision level is based on the memory level plus an
architecturally provided, fixed, decision procedure.
 The decision level proceeds in a two phase elaborate­decide
cycle.
 During elaboration, the memory is accessed repeatedly, in
parallel, until quiescence is reached; that is, until no more
productions can execute.
 This results in the retrieval into working memory of all of
the accessible knowledge that is relevant to the current
decision.
 After quiescence has occurred, the decision procedure
selects one of the retrieved actions based on the preferences
that were retrieved into working memory.

44. Goal Level
 A general intelligence must be able to set and work
towards goals.This level is based on the decision level.
 Goals are set whenever a decision cannot be made; that is,
when the decision procedure reaches an impasse.
 Impasses occur when there are no alternatives that can be
selected (no­change and rejection impasses) or when there
are multiple alternatives that can be selected, but
insufficient discriminating preferences exist to allow a
choice to be made among them (tie and conflict impasses).

45. Impasse Resolution
 Whenever an impasse occurs, the architecture generates
the goal of resolving the impasse which becomes the
subgoal.
 Along with this goal, a new performance context is created.
 The creation of a new context allows decisions to continue
to be made in the service of achieving the goal of “resolving
the impasse”.
 A stack of impasses is possible.
 The original goal is resumed after all the impasse stack is

46. Learning through Chunking
 In addition to all above levels, a general intelligence
requires the ability to learn.
 All learning occurs by the acquisition of chunks­­
productions that summarize the problem solving that
occurs in subgoals, a mechanism called “Chunking”
 The actions of a chunk represent the knowledge generated
during the subgoal; that is, the results of the subgoal.

47. Evolution of Soar
YEAR VERSION IMPLEMENTED IN
1982 Soar 1 Lisp
1983 Soar 2 Lisp/OPS5
1984 Soar 3
1986 Soar 4
1989 Soar 5
1992 Soar 6 C
1996 Soar 7 Tcl/tk
1999 Soar 8 SGIO

48. Soar 9 : Interesting Developement
 Unifying Cognitive Functions and Emotional Appraisal
 The functional and computational role of emotion is open to
debate.
 Appraisal theory is the idea that emotions are extracted
from our evaluations (appraisals) of events that cause
specific reactions in different people.
 The main controversy surrounding these theories argues
that emotions cannot happen without physiological arousal.

49. Appraisal's Detector
 This theory proposes that an agent continually evaluates a
situation and that evaluation leads to emotion.
 The evaluation is hypothesized to take place along multiple
dimensions, such as
­ goal relevance
­ goal conduciveness
­ causality and control
 These dimensions are exactly what an intelligent agent
needs to compute as it pursues its goals while interacting
with an environment.

50. Conclusion
 This field still has far to travel before we understand fully
the space of cognitive architectures and the principles that
underlie their successful design and utilization.
 However, we now have over two decades’ experience with
constructing and using a variety such architectures for a
wide range of problems, along with a number of challenges
that have arisen in this pursuit.
 If the scenery revealed by these initial steps are any
indication, the journey ahead promises even more
interesting and intriguing sites and attractions.

51. Soar 9 : Appraisal Detector

52. References
1) SOAR : An Architecture for General Intelligence, John E.
Laird, Allen Newell, Paul S. Rosenbloom,1986.
2) A preliminary analysis of the Soar architecture as a basis
for general intelligence, John E. Laird, Allen Newell, Paul
S. Rosenbloom, 1989.
3) http://en.wikipedia.org/wiki/Cognitive_architecture
4) http://cs.gmu.edu/~eclab/research.html
5) http://en.wikipedia.org/wiki/Unified_theory_of_cognition
6) http://cll.stanford.edu/research/ongoing/icarus/
7) http://en.wikipedia.org/wiki/Artificial_consciousness
8) http://plato.stanford.edu/entries/functionalism/

53. References
9) A Survey of Cognitive Architectures, David E. Kieras,
University of Michigan .
10) Connectionism and Cognitive Architecture : A Critical
Analysis, Jerry A. Fodor and Zenon W. Pylyshyn, Rutgers
Center for Cognitive Science, Rutgers University, New
Brunswick, NJ.
11) Human Cognitive Architecture, John Sweller, University
of New South Wales, Sydney, Australia.
12) http://cogarch.org/index.php/Soar/Architecture
13) http://code.google.com/p/soar/wiki/Documentation

54. References
14) A Gentle Introduction to Soar : An Architecture for
Human Cognition : 2006 Update, Jill Fain Lehman, John
Laird,Paul Rosenbloom.
15) http://sitemaker.umich.edu/soar/home