Kurso temos

1
Kurso temos

• Ivadas. Kurso esme, tikslai, reikalavimai.
• Painimo aikinimo istorija (hylemorfizmo ir
  intencionalumo savoku atvilgiu).
• Bendriausia painimo reikiniu klasifikacija.
  Reprezentacinis painimo aikinimas.
• Reprezentacinio painimo saranga. Kognityvines
  architekturos samprata, tipai.
• Painimas ir technine sistema. Simbolinis
  painimas. Turingo maina. Turingo testas. Von
  Neumanno architektura. Nuoseklus (sutelktas)
  informacijos apdorojimo principas. Dirbtinis
  intelektas.
• Konekcionizmas. Neurokompiuteriai. Lygiagretus,
  paskirstytas informacijos apdorojimo principas.
  Ne-Von Neumanno architektura.
• Tiesioginis (nereprezentacinis), analoginis
  painimas. Analoginiai kompiuteriai.
• Painimas ir veiksmas agentai, robotai.

2
Painimas ir veiksmas agentai, robotai
Degalines robotas
okantis robotas
Robotas humanoidas
Sony Aibo unys
Surinkimo robotas
Robotas humanoidas
3
Painimas ir veiksmas
4
Veiksmas, veikejas, elgesys

• Lot. agere, veikti. I to agentas, veikejas
• Lot. act, veiksmas.
• Angl. action
• Angl., behavior - elgesys

5
Agento galimybes
6
Intelektualus agentas

• Kompiuteriu moksle tai programinis agentas
  (software agent), turintis tam tikra dirbtinio
  intelekto pavidala
• Smulkiau
• While the working of software agents used for
  operator assistance or data mining (sometimes
  referred to as bots), are often based on fixed
  pre-programmed rules, "intelligent" here implies
  the ability to adapt and learn.

7
Robotas -

• - tai fizine sistema autonomikai jaucianti
  aplinka ir joje veikianti
• Robot vardas pagal Karel Capek

8
More definitions of Robots

• an active, artificial agent whose environment is
  the physical world. Russell and Norvig
• An intelligent robot is a mechanical creature
  which can function autonomously Murphy
• An intelligent robot is a machine able to
  extract information from its environment and use
  knowledge about its world to move safely in a
  meaningful and purposive manner. - Arkin

9
Robot timeline
Isaac Asimovs Laws of Robotics
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.
Karl Capek
Rossums Universal Robots
...
I, Robot
1921
1950
10
Why Robotics?
pumping gas
assembly
Practice
welding
dancing
Promise
eating cars
packaging
11
Why Robotics?
kismet
Sony Aibo dogs
Vibrant field
other competitions
Harold Cohens Aaron
12
Robotics a synthetic field

• Biology
• Neuroscience
• Cognitive Science

• Mechanical Eng.
• Electrical Eng.
• Control

• CS AI
• Vision, HCI
• Software Dev.

Robotics

• Architecture
• Mapping

• Sensing and Motion
• Cooperative Robotics

• Visual Control
• Manipulation

• Planning / Modeling

• Robot Learning

13
Robotu architektura
14
Architekturu tipai (Approaches to Hierarchy)

• Laiko (Temporal
• Elgesio (Behavioral
• Funkcines (Functional

15
What is a Robot Architecture?

• Conceptual Framework for Designing Robot Systems
• Implementation Glue for Integrating and
  Coordinating Robot Systems
• Important Constraints on Architecture Design
• Situated / embedded / interacting
• High perceptual bandwidth
• Dynamic / unpredictable environment
• Uncertainty in sensing and action

16
Reaktyvines architekturos

• Justi - veikti

17
Paprasto refleksinio agento architektura

• Simple reflex agent http//en.wikipedia.org/wiki/I
  ntelligent_agent
• Sense-act

18
Sudetingesnis atvejis

• Justi- planuoti- veikti (sense-plan-act)

19
Besimokancio agento architektura
20
Priminimas Dekarto koordinaciu
sistemoshylemorfine interpretacija(pagal
Aristotelio formos ir materijos perskyra)

21
Agento procesu interpretacija

• Jutimas (formos peremimas). Pagal Aristoteli,
  jutimo procese i dalyko atimama (perimama) jo
  forma. Tai vadinama abstrakcija.
• Veikimas (ikunijimas, materijos suteikimas)

22
Paprasto refleksinio agento architekturos
vaizdavimas hylemorfineje koordinavimo sistemoje
Agentas
Aplinka

Jutimas (formos peremimas)
i perskyra ignoruojama
Veikimas (ikunijimas)
23
Sudetingo refleksinio agento architekturos
vaizdavimas hylemorfineje koordinavimo sistemoje
Agentas
Aplinka

Jutimas (formos peremimas)
planavimas
Veikimas (ikunijimas)
24
Hierarchine roboto architektura
25
Rodney Brooks pasiulyta roboto elgesio
architektura (itraukianti, apimanti - Subsumption
Architecture)
navigate behavior
wander behavior
runaway behavior
FSM / DFA
26
Vardai robotikoje

• Rodney Brooks, is Director of the MIT Computer
  Science and Artificial Intelligence Laboratory,
  and is the Panasonic Professor of Robotics. He is
  also co-founder and Chief Technical Officer of
  iRobot Corp (nasdaq IRBT).

27

• Subsumption architectures, developed e.g. by
  Rodney Brooks (though it could be argued whether
  they are cognitive). http//en.wikipedia.org/wiki/
  Cognitive_architecture
• Brooks has argued strongly against symbolic
  processing approaches to creating intelligent
  machines, which had been the focus of AI since
  the days of Alan Turing, directly tracing back to
  the work of Gottlob Frege. Instead, Brooks has
  focused on biologically-inspired robotic
  architectures (e.g., the Subsumption
  architecture) that address basic perceptual and
  sensorimotor tasks. These had been largely
  dismissed as uninteresting by the mainstream AI
  community, which was far more interested in
  reasoning about the real world than in
  interacting with it. Conversely, Brooks argued
  that interacting with the physical world is far
  more difficult than symbolically reasoning about
  it. This perspective is perhaps best and most
  eloquently described in his classic paper,
  Elephants Don't Play Chess.

28
Robotu architekturos itakos Uexkuelio
biosemiotika
29
Jakob von Uexkuel(1864-1944)

• Born in Keblaste (now Mihkli), Estonia. He
  studied zoology in the University of Tartu (then
  Dorpat).
• Sebeok, Thure von Uexkuel (son) discovered him
  as semiotician
• A representative of biological semiotics, a
  theoretical biologist.
• http//home.datacomm.ch/biografien/biografien/uexk
  uell.htm
• Jakob von Uexküll Centre. TARTU, ESTONIA (founded
  in 1993). http//www.zbi.ee/uexkull/
• He is recognised as the founder of biosemiotics
  (Theory of meaning, Bedeutungslehre, 1940).
• Biosemiotics and theoretical biology links
  (papers online) http//www.zbi.ee/kalevi/vaheleh
  t.htm
• Links http//www.zbi.ee/uexkull/link.htm
• In wikipedia. http//en.wikipedia.org/wiki/Jakob_v
  on_UexkC3BCll
• K.Kull (2001). Jakob von Uexküll An
  introduction. Semiotica 134(1/4) 1-59.

30
Semiotics of Jakob von Uexkuel main ideas

• Umwelt. Umwelt is the semiotic world of
  organism. It includes all the meaningful aspects
  of the world for a particular organism. Thus,
  Umwelt is a term uniting all the semiotic
  processes of an organism into a whole. Kalevi
  Kull. On semiosis, Umwelt, and semiosphere.
  http//www.zbi.ee/kalevi/jesphohp.htm
• Umwelt is a subjective world of an organism. It
  consist of the perceptual (Merkwelt) and
  operational parts (Wirkwelt).
• Functional circle. It describes interaction of an
  organism with its Umwelt.

31
Functional circle of Uexkuell

• Functional describes interaction of an organism
  with its Umwelt.
• Receiver of meaning (subject) lt - gt Carrier of
  meaning (object)
• Carrier of perceptual sygnal, perceptual world,
  perceptual organ
• Carrier of effect sygnal, operational (motor)
  world, effector organ
• In Biologische Semiotik. http//homepage.ruhr-uni
  -bochum.de/Udo.Figge/texte/uexkuell.htm

32
Literatura

• Action. http//plato.stanford.edu/entries/action/
• ACTION THEORY PAGE. http//www.angelfire.com/ab3/
  freewill/ActionTheory.html
• Rodney A. Brooks. Intelligence Without Reason.
  MASSACHUSETTS INSTITUTE OF TECHNOLOGY.
  ARTIFICIAL INTELLIGENCE LABORATORY. A.I. Memo No.
  1293 April, 1991. Prepared for Computers and
  Thought, IJCAI-91
• Rodney A. Brooks. Elephants Don't Play Chess. MIT
  Artificial Intelligence Laboratory, Cambridge, MA
  02139, USA Robotics and Autonomous Systems 6
  (1990) 3-15.http//people.csail.mit.edu/brooks/pap
  ers/elephants.pdf Brooks has argued strongly
  against symbolic processing approaches to
  creating intelligent machines, which had been the
  focus of AI since the days of Alan Turing,
  directly tracing back to the work of Gottlob
  Frege. Instead, Brooks has focused on
  biologically-inspired robotic architectures
  (e.g., the Subsumption architecture) that address
  basic perceptual and sensorimotor tasks. These
  had been largely dismissed as uninteresting by
  the mainstream AI community, which was far more
  interested in reasoning about the real world than
  in interacting with it. Conversely, Brooks argued
  that interacting with the physical world is far
  more difficult than symbolically reasoning about
  it. This perspective is perhaps best and most
  eloquently described in his classic paper,
  Elephants Don't Play Chess.
• Kalevi Kull. Biosemiotics in the twentieth
  century a view from biology. Semiotica vol.
  127(1/4), pp. 385-414 (1999). http//www.zbi.ee/7
  Ekalevi/bsxxfin.htm

33
Priminimas Reprezentacinio painimo ribos

• Pateikta dvipuse formaliojo ir materialiojo
  painimo schema aikina reprezentacini painima,
  tai yra, toki, kai objektas paistamas per jo
  formos ir/arba materijos reprezentacija
• Pastaba materialusis reprezentacinis painimas
  neretai tapatinamas su nereprezentaciniu, tai
  yra, pripaistamos tik formaliosios
  reprezentacijos
• Priminimas alternatyva visam reprezentaciniam
  painimui yra nereprezentacinis painimas
• Nereprezentacinis painimas yra keliu tipu
• Analoginis (tiesioginis, betarpikas painimas,
  kai tarp subjekto ir objekto nera tarpines
  grandies, reprezentacijos)
• Ikunytas, pvz., per veiksma, situacinis
  (priklausoma nuo aplinkos) painimas
• Paskirstytas
• hetero-reprezentacinis (kai vienalyte
  reprezentacija neimanoma, pvz., tokiu dalyku kaip
  kentauras), iracionaliosios reprezentacijos.
• P.s. Anksciau apibudinta schema netinka
  tiesioginio painimo aikinimui

34
Kognityvines architekturos
35
Architektura

• Tai sandara, saranga visumos ir daliu santykis

36
Kognityvines architekturos keturi tipai

• Simboline architektura, von Neumano, skaitmeninio
  kompiuterio
• Subsimboline architektura (dirbtinio
  neurotinklo), ne von Neumano, paskirstyto ir
  lygiagretaus informacijos apdorojimo
• Analogine architektura (analoginio kompiuterio)
• Agento (veiksmo) architektura (roboto, situacine,
  susieta su aplinka, ikunyta)
• Ar yra daugiau (pamatiniu) architekturos tipu?

37
Kognityvines architektura ir reprezentacija

• Simboline architektura tipinis reprezentacinio
  painimo atvejis
• Subsimboline architektura (dirbtinis
  neurotinklas) sutelktu reprezentaciju nebuvimas
  (paskirstytos reprezentacijos)
• Analogine architektura (analoginis kompiuteris),
  tiesioginis painimas, be reprezentacijos
• Agento (veiksmo) architektura (roboto, situacine,
  susieta su aplinka, ikunyta)

38
Architekturu lyginimo budas variantas
39
Priminimas Dekarto koordinaciu
sistemoshylemorfine interpretacija(pagal
Aristotelio formos ir materijos perskyra)

40
Pabandyti visas architekturas patalpinti
hylemorfineje sistemoje