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
Artificial Cognition

• What is the relationship between a and b?
• a b

3
(No Transcript)
4
Science-fiction v Science-fact

• Embodied cognition most commonly in a robot
• 1921 Robot,
• One of the mechanical men and women in Capek's
  play hence, a machine (sometimes resembling a
  human being in appearance) designed to function
  in place of a living agent, esp. one which
  carries out a variety of tasks automatically or
  with a minimum of external impulse.
• 1923Mr. G. Bernard Shaw defined Robots as
  persons all of whose activities were imposed on
  them.
• Oxford English Dictionary

5
Asimovs robots

• "In the 1920's a robot .... Under the influence
  of the well-known deeds and ultimate fate of
  Frankenstein and Rossum, - robots were created
  and destroyed their creator .... My robots were
  machines designed by engineers, not pseudo-men
  created by blasphemers"
• First Law
• A robot may not injure a human being, or,
  through inaction, allow a human being to come to
  harm.
• Second Law
• A robot must obey orders given it by human
  beings, except where such orders would conflict
  with the First Law.
• Third Law
• A robot must protect its own existence as long
  as such protection does not conflict with the
  First or Second Law.
• From Runaround1942

6
Data

• Data, was created in an attempt to bring
  "Asimov's dream of a positronic robot" to life.
  Asimov's robots are "positronic" because
  positrons had just been discovered when he
  started writing robot stories
• (Handbook of Robotics, 56th Edition)
• Data has 100,000 terabytes of memory he had a
  storage capacity of 800 quadrillion bits (100
  quadrillion bytes). The storage capacity of the
  human brain to approximately 3 teraBITS, which
  would mean that Data's brain could contain
  everything from over 260,000 human brains.

7
Positronic Networks

• The system uses the decay of positrons to form a
  sophisticated neural network
• Dr. Noonian Soong, 2338
• Is Data a sentient being?
• How easy is it to judge
• Intelligence.
• Self-awareness.
• Consciousness ?

8
Strong AI v Weak AI

• AI has a philosophical theory of mind
• The theory is that human cognitive mental states
  can be duplicated using computing machinery
  (electronic or biological)
• A distinction is made between two levels of
  duplication
• Strong AI machines can be made to think and have
  genuine understanding and other cognitive states
• Weak AI machines can simulate thought but does
  not claim that they will understand what is being
  reasoned

9
Strong AI v Weak AI

• Searles Chinese room argues against strong AI as
  a machine can manipulate symbols without
  understanding.
• A machine a process syntax but not the semantics.
• Arguments against Searle include taking the
  whole system to be intelligent and robot response
  that no understanding can exist without the
  sensory connections to the world and related
  objects
• Charmers admit that consciousness is really
  impossible to define.
• Sloman believes that the discussion of
  consciousness is a real mess

10
AI

• http//aimovie.warnerbros.com/

11
Emotions and Love in Robots

• http//www.cloudmakers.org/

12
Robot Ethics

• Robot Code of Ethics to Prevent Android Abuse,
  Protect Humans
• National Geographic News Stefan Lovgren
• March 16, 2007
• The government of South Korea is drawing up a
  code of ethics to prevent human abuse of
  robotsand vice versa.
• The so-called Robot Ethics Charter will cover
  standards for robotics users and manufacturers,
  as well as guidelines on ethical standards to be
  programmed into robots, South Korea's Ministry of
  Commerce, Industry and Energy announced
• 8th International Conference on Ethics Across the
  Curriculum
• Dartmouth College, New Hampshire November 16-19,
  2006
• Robot Morals and Human Ethics Wendell Wallach
  (Yale University)
• gtresearchnews.gatech.edu/reshor/rh-w07/robot-ethi
  cs.pdf
• http//www.roboethics.org/site/

13
Robot Ethics

• Hiroshi Ishiguru of Osaka University is the
  co-creator of Repliee Q1 and Q2, a female
  android.
• It's too soon to dictate the future of robot
  ethics based on Asimov's laws.
• If we have a more intelligent vehicle, who takes
  the responsibility when it has an accident?
• We can ask the same question of a robot. Robots
  do not have human-level intelligence. It is
  rather similar to a vehicle today."

14
Emotional Robotics

• Dr Susan Calvin obtained her bachelor's degree at
  Columbia in 2003 and began graduate work in
  cybernetics.
• Asimov (1940)

15
Emotions and Love in Robots

• The question is not whether intelligent machines
  can have any emotions, but whether machines can
  be intelligent without emotions.
• Marvin Minsky
• In the book Speaker for the Dead science-fiction
  author Orson Scott Card suggested that when such
  an intelligence will emerge, it will be scared to
  let us know of its existence, knowing that some
  people, fearing anything nonhuman, will do their
  best to terminate it.

16
Existing Emotional Architectures

• ISAC (Kawamura 06) uses Haikonens System
  Reactions Theory of Emotions
• Sensations - Reactions
• Amygdala model (John Taylor 2006)
• ASD Action selection dynamics (Maes 1990)
• ALEC Asynchronous Learning by Emotions and
  Cognition (Gadanho 2002)
• DARE (Marcia et al. 2001)
• AD with ECS emotional control system (Malfaz et
  al.)
• http//www.mindraces.org/documents/deliverable/del
  iverable5/D5_1.pdf

17
Emotional Architecture
18
Creating Emotions in a Robot
Visible Emotions shown useful in human-robot
interaction and should speed up robot-robot
interaction However these can be top-down
interpretations of internal deterministic
states - Internal Emotions emerge from embodiment
and interaction with an environment Motives
(goals) interact with beliefs (predictions) to
produce emotions. Feelings arise after enaction
of emotions. Reactive mechanisms make
possible alarm-driven primary emotions while
Deliberative mechanisms make possible
secondary emotions using global alarm mechanisms
linked to deliberative processes. (Sloman)
19
Robotic Emotions

• Robots need real emotions to successfully
  complete complex real-world tasks.
• Emotions can set goals balance explore vs
  exploit
• Emotions can modify existing behaviours
• Emotions facilitate action in unknown domains