Lecture 3 Embodiment: Concept and Models Fabio Bonsignorio The - - PowerPoint PPT Presentation

▶
lecture 3 embodiment concept and models
SMART_READER_LITE
LIVE PREVIEW

Lecture 3 Embodiment: Concept and Models Fabio Bonsignorio The - - PowerPoint PPT Presentation

Jan 07, 2023 234 likes •751 views

Lecture 3 Embodiment: Concept and Models Fabio Bonsignorio The BioRobotics Institute, SSSA, Pisa, Italy and Heron Robots Todays topics short recap The classical approach: Cognition as computation Successes and failures of the

slide-1
SLIDE 1
slide-2
SLIDE 2

Lecture 3 Embodiment: Concept and Models

Fabio Bonsignorio The BioRobotics Institute, SSSA, Pisa, Italy and Heron Robots

slide-3
SLIDE 3

Today’s topics

  • short recap
  • The classical approach: Cognition as

computation

  • Successes and failures of the classical

approach

  • Some problems of the classical approach
  • The need for an embodied approach

3

slide-4
SLIDE 4

Today’s topics

  • short recap
  • The classical approach: Cognition as

computation

  • Successes and failures of the classical

approach

  • Some problems of the classical approach
  • The need for an embodied approach

4

slide-5
SLIDE 5

Today’s topics

  • short recap
  • The classical approach: Cognition as

computation

  • Successes and failures of the classical

approach

  • Some problems of the classical approach
  • The need for an embodied approach

5

slide-6
SLIDE 6

“Birth” of AI, 1956

6

Herbert Simon and Allen Newell The “Logic Theorist” Noam Chomsky, Linguist “Syntactic Structures” George A. Miller, Psychologist “The Magical Number Seven Plus or Minus Two” John McCarthy, Computer Scientist Initiator of Artificial Intelligence

slide-7
SLIDE 7

Turing Machine (1)

7

slide-8
SLIDE 8

Turing Machine (2)

8

input from tape 1 2 state of read/write head

_ _R2 HALT A AL1 BR2 B BL1 AR2 C CL1 CR2

write on tape next state of r/w head move tape L/R

slide-9
SLIDE 9

9

input from tape 1 2 state of read/write head

_ _R2 HALT A AL1 BR2 B BL1 AR2 C CL1 CR2

write on tape next state of r/w head move tape L/R

initial situation: state r/w head = 1 initial content of tape:

r/w head initial pos.

. . . A A B A A C C C C A B A C C C C B B A B . . .

slide-10
SLIDE 10

10

input from tape 1 2 state of read/write head

_ _R2 HALT A AL1 BR2 B BL1 AR2 C CL1 CR2

write on tape next state of r/w head move tape L/R

initial situation: state r/w head = 1 initial content of tape:

r/w head initial pos.

. . A A B A A C C C C A B A C C C C B B A B . . .Turing

Machine (4)

The Universal Turing Machine

slide-11
SLIDE 11

Turing Machine (5)

11

Cartoon by
 Roger Penrose an “embodied” Turing Machine

slide-12
SLIDE 12

Functionalism and the

“Physical Symbol Systems

12

biological electronic mechanical Swiss cheese Hilary Putnam (American Philosopher)

slide-13
SLIDE 13

Functionalism and the

“Physical Symbol Systems

Model/Representation:

13

slide-14
SLIDE 14

GOFAI

G O F A I

14

slide-15
SLIDE 15

Classical AI: Research areas

  • problem solving
  • knowledge representation and reasoning
  • acting logically
  • uncertain knowledge and reasoning
  • learning and memory
  • communicating, perceiving and acting
  • (adapted from Russell/Norvig: Artificial intelligence, a modern

approach)

15

slide-16
SLIDE 16

Today’s topics

  • short recap
  • The classical approach: Cognition as

computation

  • Successes and failures of the classical

approach

  • Some problems of the classical approach
  • The need for an embodied approach

16

slide-17
SLIDE 17

Classical AI: Successes

  • search engines
  • formal games (chess!)
  • text processing systems/translation —> next

week

  • data mining systems
  • restricted natural language systems
  • appliances

17

Indistinguishable from computer applications in general

slide-18
SLIDE 18

Chess: New York, 1997

  • 18

1 win 2 wins 3 draws

slide-19
SLIDE 19

Classical AI: Failures

  • recognizing a face in the crowd
  • vision/perception in the real world
  • common sense
  • movement, manipulation of objects
  • walking, running, swimming, flying
  • speech (everyday natural language)

19

in general: more natural forms of intelligence

slide-20
SLIDE 20

Why is perception hard?

20

slide-21
SLIDE 21

Today’s topics

  • short recap
  • The classical approach: Cognition as

computation

  • Successes and failures of the classical

approach

  • Some problems of the classical approach
  • The need for an embodied approach

21

slide-22
SLIDE 22

Fundamental problems of the classical approach

22

real world virtual, formal world

Monika Seps, chess master former master student AI Lab, Zurich

in general: anything to do with real world interaction fundamental differences: real — virtual

slide-23
SLIDE 23

Fundamental problems of the classical approach

23

real world virtual, formal world

in general: anything to do with real world interaction fundamental differences: real — virtual

slide-24
SLIDE 24

Differences real vs. virtual worlds

24




slide-25
SLIDE 25

Successes and failures of the classical approach

25

successes applications (e.g. Google) chess manufacturing (applications:“controll ed”artificial worlds) failures foundations of behavior natural forms of intelligence interaction with real world

slide-26
SLIDE 26

Industrial environments vs.

industrial environments environment 
 well-known little uncertainty predictability (“controlled”artificial worlds) real world environment limited knowledge and predictability rapidly changing high-level of uncertainty

26

slide-27
SLIDE 27

Industrial robots vs. natural systems

27

principles:

  • strong, precise, fast motors
  • centralized control
  • computing power
  • optimization

Industrial robots

slide-28
SLIDE 28

Industrial robots vs. natural systems

28

principles:

  • low precision
  • compliant
  • reactive
  • coping with

uncertainty

human s

no direct transfer of methods

slide-29
SLIDE 29

Fundamental problems of classical approach

  • “symbol grounding problem”
  • “frame problem”
  • “homunculus problem”

29

slide-30
SLIDE 30

The “symbol grounding” problem

real world:
 doesn’t come
 with labels ...

30

Gary Larson

slide-31
SLIDE 31

The “frame problem” Maintaining model of real

  • the more detailed


the harder

  • information 


acquisition

  • most changes:


irrelevant to current
 situation

31

slide-32
SLIDE 32

Today’s topics

  • short recap
  • The classical approach: Cognition as

computation

  • Successes and failures of the classical

approach

  • Some problems of the classical approach
  • The need for an embodied approach

32

slide-33
SLIDE 33

33

Two views of intelligence

classical: 
 cognition as computation embodiment: 
 cognition emergent from sensory-motor and interaction processes

slide-34
SLIDE 34

The need for an embodied perspective

34

  • “failures” of classical AI
  • fundamental problems of classical

approach

  • Wolpert’s quote:
slide-35
SLIDE 35

The need for an embodied perspective

35

“Why do plants not have brains?”

slide-36
SLIDE 36

The need for an embodied perspective

36

“Why do plants not have brains? The answer is actually quite simple — they don’t have to move.” Lewis Wolpert, UCL evolutionary perspective on development of intelligence/cognition

slide-37
SLIDE 37

The need for an embodied perspective

37

  • “failures” of classical AI
  • fundamental problems of classical

approach

  • Wolpert’s quote: Why do plants not …?
  • Interaction with environment: always

mediated by body

slide-38
SLIDE 38

Today’s topics

  • short recap
  • The classical approach: Cognition as

computation

  • Successes and failures of the classical

approach

  • Some problems of the classical approach
  • The need for an embodied approach

38

slide-39
SLIDE 39

The “frame-of-reference” problem — introduction

39

Video “Heider and Simmel”

slide-40
SLIDE 40

The “frame-of-reference” problem — introduction

40

Video “Heider and Simmel”




slide-41
SLIDE 41

“Frame-of-reference” Simon’s ant on the beach

41

slide-42
SLIDE 42

“Frame-of-reference” Simon’s ant on the beach

42

food nest

slide-43
SLIDE 43

“Frame-of-reference” Simon’s ant on the beach

  • simple behavioral rules
  • complexity in interaction, 


not — necessarily — in brain

  • thought experiment:


increase body by factor of 1000


43

slide-44
SLIDE 44

“Frame-of-reference” Simon’s ant on the beach

44

food nest

new path?

slide-45
SLIDE 45

“Frame-of-reference” F-O-R

  • perspectives issue
  • behavior vs. mechanism issue
  • complexity issue

45

slide-46
SLIDE 46

“Frame-of-reference” F-O-R

  • perspectives issue
  • behavior vs. mechanism issue
  • complexity issue

46

slide-47
SLIDE 47

Intelligence:

Hard to agree on definitions, arguments

  • necessary and sufficient conditions?
  • are robots, ants, humans intelligent?

more productive question:

“Given a behavior of interest, how to implement it?”

47

slide-48
SLIDE 48

Communication through interaction with

  • exploitation of interaction with environment

simpler neural circuits

48

angle sensors in joints

“parallel, loosely coupled processes”

slide-49
SLIDE 49

Emergence of behavior: the quadruped “Puppy”

49

  • simple control (oscillations of 


“hip” joints)

  • spring-like material properties 


(“under-actuated” system)

  • self-stabilization, no sensors
  • “outsourcing” of functionality

morphological computation

actuated:


  • scillation


springs
 passive


slide-50
SLIDE 50

Implications of embodiment

50

Pfeifer et al.,Science, 16 Nov. 2007

“Puppy”, But Also Cruse

slide-51
SLIDE 51

Implications of embodiment

51

Pfeifer et al.,Science, 16 Nov. 2007

“Puppy” which part of diagram is relevant? 
 —>