Further Cognitive Systems

1
Further Cognitive Systems

• Learning
• Environmental interaction
• Artificial cognition?
• Current cognitive systems
• Science-fiction v fact
• Architectures
• Perception, Representation, Reasoning, Learning
  Action
• Learning Cognitive Systems
• Problems in LCS
• Advances in LCS

2
Current Cognitive Systems

• When did cognitive systems start?
• Philosophically
• Prior to Plato (427-347 BC)
• Computationally
• Coincided with large computers.
• E.g. IBM-704 (1956)

3
Early Examples

• W. Pitts and W. S. McCulloch, "How we know
  universals," Bull. Math. Biophys., vol. 9, pp.
  127-147, 1947.
• D. O. Hebb, "The Organization of Behavior," John
  Wiley and Sons, Inc., New York N. Y.,
  1949
• A. L. Samuel, "Some studies in machine learning
  using the game of checkers," IBM J. Res. Dev.,
  vol. 3,pp. 211-219, July 1959.
• F. Rosenblatt, "The Perceptron" Cornell
  Aeronautical Lab. Inc., Ithaca, N. Y. Rept.
  VG-1196, January 1958

4
Symbolic Intelligence

• Physical symbol system
• Any facet of human intelligence can be
  understood and described precisely enough for
  machine to simulate it,
• The representation uses symbolic structures
• E.g. Expert Systems
• (DENDRAL, 1965 Feigenbaum and Lindsay)
• 1. IF the engine will not turn over
• AND the lights do not come on
• THEN the battery is dead
• 2. IF the battery is dead
• THEN the car will not start

5
Connectionist Systems

• Connectionist AI
• Computing elements resemble an abstraction of
  our own neural circuitry

6
Traditional v Modern

• GOFAI (good old fashioned AI)
• Logical and psychological models
• Case Based Reasoning
• Expert systems
• NEWFAI
• Biologically Inspired
• Evolutionary Computation
• Semantic webs
• Neural Networks

7
Cognitive Systems v AI

• AI is the science of making machines do things
  that would require intelligence if done by
  humans
• Marvin Minsky
• Cognitive Systems
• "Cognitive systems are natural or artificial
  information processing systems, including those
  responsible for perception, learning, reasoning
  and decision-making and for communication and
  action"
• DTI Foresight initiative
• Perception and Action embody intelligence

8
Perception

• Through senses
• Sight - Vision systems
• Hear - Speech recognition
• Smell - Olfaction
• Touch - Haptics
• Taste - ?
• ?
• Telepathy
• ESP

9
Action

• Robot embodiment

10
Cogric

• Cognitive robotics, intelligence and control
• 16-18 August 2006
• http//www.cogric.reading.ac.uk/

11
Owen Holland University of Essex

• How could the agent achieve its task (or
  mission)?
• by being preprogrammed for every possible
  contingency? No
• by having learned the consequences for the
  achievement of the mission of every possible
  action in every contingency? No
• by having learned enough to be able to predict
  the consequences of tried and untried actions, by
  being able to evaluate those consequences for
  their likely contribution to the mission, and by
  selecting a relatively good course of action?
  Maybe

12
Information Measures
Quantifying Information in Networks The Problem
13
standard view
J Kevin ORegan Laboratoire Psychologie de la
Perception Centre National de la Recherche
Scientifique Université René Descartes - Paris
5
Explanatory gap!
14
Sensation exercising a skill
No more explanatory gap!
15
Biological reflection properties
R
LMSr
LMSi

• for a biological organism
• reflection properties are constraints over
  sensory inputs
• set of reflection properties is finite
  dimensional
• ? finite number of singular reflection properties

16
D. Philipona J K ORegan, 2006
17
CornellMIT Delft
Rolf Pfeiffer University of Zurich
Passive Dynamic Walker (Cornell)
mehr später
Denise (Delft)
Qrio (Sony)
Asimo (Honda)
18
Puppy on the treadmill
Rolf Pfeiffer University of Zurich
19
Engage in a Behavioral Task And Adapt Behavior
When An Important Environmental Event Occurs
Jeff Krichmar The Neurosciences Institute
20
Allow Comparisons with Experimental Data Acquired
from Animal Systems
Jeff Krichmar The Neurosciences Institute
ECout
CA1
ECin
DG
CA3
21
Common Topics

• Morphology
• Information theoretic
• Attention
• Working memory
• Emotions
• Embody
• Robotics as tools and/or platforms
• Feedback/feedforward

22
Common Topics

• Consciousness
• Variety of representations
• Learning and development
• Interaction and Communication
• Structural properties of neural systems

23
Concluding Remarks

• Dont make cognition hard for ourselves
• Models are useful,
• but the mind is not so clear-cut
• Human cognition is a good model,
• but desired behaviour may be achieved by other
  models
• Increasingly powerful tools assist in advancing
  cognitive robotics, e.g., computational power,
  engineering materials and neurological
  understanding.