0% found this document useful (0 votes)
37 views24 pagesMujtaba-Artificial Intelligence
The document discusses the concept of intelligent agents that can perceive their environment through sensors and perform actions based on that information. It outlines various types of agents, including rational agents and goal-based agents, and their decision-making processes. Additionally, it touches on machine learning, cognitive abilities, and the implications of artificial intelligence in mimicking human thought processes.
Uploaded by
manCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
0% found this document useful (0 votes)
37 views24 pagesMujtaba-Artificial Intelligence
The document discusses the concept of intelligent agents that can perceive their environment through sensors and perform actions based on that information. It outlines various types of agents, including rational agents and goal-based agents, and their decision-making processes. Additionally, it touches on machine learning, cognitive abilities, and the implications of artificial intelligence in mimicking human thought processes.
Uploaded by
manCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
AnkiSciad Intelligence
Nv
Prgent
=> fn agent- tsa bem that __can .
_Petcetye its envivonment thvough _sensass__
and act upon that | envisenmentt Ahteah.
oN Z
capabilities +o __gevfotm ackions+
(Needs sienflar Sneut #4 oot dex'cos,)
X p 2. 2 A
KO Ache
Trough which we cant
petlotn actions.
Agents Hore." |
k Ervivonment st Tesk dependent
‘
=y Fetteption: rr that Beatin CHeanings
Foun as -PekeepHion task.
=> Sensors sep, Boch»
A Sensor: only
Q
Ed lees agents . can 1 _petfonn
u
an Rational Action
a “shoul ee an action fon hee fox (a- cnet vst
Sate te oktd “arosledge) » Rationabis ft
As: Detect: from Ant (compel
diferent) +o
Infinite knowledge sieiage. —
ComMporests Nereyx ests So, we _
Bact called Yattonal agents
ommiscionce:.:
Tnbintte. Enowlelae Never exists:
S Dedsfon mating. biftiestt ;
ven _ move. Ynawledge. As_pesenti—
=> CTatme-taliing\.
ses Performance paler max
| enouledge- (also be a bottle + bet
now mn 4
» Aoerts. Con _perStmn actions tn cider
to modify fire pefcep is Sr
Ag. to cktnla vsefl ‘heal
Pn a aa ;
Poor is "S henteel o tS aon.
aren us thi abit 40. fen tt adapt -., cb Cowect Efhe?, 1s
Beton’ should begat co ‘
co — i ——
i) Envivonment .
Shovtd be
nts ae
ed {For whith -etltanchen one
e Sack, Liv. t
Sensors ALN ty) Aetuator:
to ewYonment: :
DIS) Yel Bxpaielcl .
a
“LEntonment
—Obsefiable.
Coins plete ‘
Yate | tabiobgs 2 jolt envitonmend 1S
Sd(Tf next & ich
hi mee :
pletely cetorninatse) <) lem t (Ovid bbe
als . erected bY GPE" 5 1 a chastic At agpeas ern enwonren: aa ne .opken to, ay
ae Cnvitonment is sip deterministic lbut PF
they hae mullipe options, while Performing 4 action
— trent will be shetegic. detemministo enuitory prety
i) [Episod vs T&apentiad 7
= se ee is _t Coternt decision
n tom C
ea Ep ode Meat ‘st could affect oth fsture
ror a Eien TF
W
‘vs ynant
ca. OBS agent
OES es
Multiple _ cers 3
No eed 40 loole at
2 ti stabi. lee aie
c ceig the selon. Cred
Agent Opetedti. Obsetie Calestytted ng, ng
Da ttsAP oi an env nent xo pen Five. ok
Of with: other Agent - competed ve
|? Semi Dycomic.: 4c ,
Eovit srment —cloesnct
change wih passage. of dee
Dot petfamoance af ¢pr does.
14 t pe iss
# Sienp le neh a
x MocleD based Retiox Age nts. ‘
ix Goal-based . Agents
qe
Agents
_* Motivational Agent's - 4). Sieaplen\ Reflex: (Base ype. functions pertoin).
a ise \ rtuatect in CANES
ein Sev EM eit neni - thisvg
SeASOIS A Ae —— 2 |
1) Fix! Conditivn utes (Present) +
pests
‘ state ofc Witenine
cthen |. bdcite al des toe
le. (At dnat ae what they
Reve to: de). 1 derstand
ne + Ealicorrt Sense ae State.
: vi Ofte usor id)
Bn decide acrar te, gto the pre-def
ich aston as fs e! Be
Hons On world -O'sluing F_We ne
a: ate eke s
On B age: _conditisn oles , sctheyl ecto
current State oft world and after
that deciding. phase. on the basis
pf baste condittans «Mey decide.
Lo ef fom actionss Flaw: (Not implement Particularly goals).
niin). @af-® What _ovttame willbe? ite action
i pill. be decfold! . after’ outcomes
qh_ qr oavit: Check, how other environment
AGorct LAW wort and wor /do action
pececling +0 +ol -
wt OM 2g. Te tac, 02.
oa basecl ageuks, -
; Lota envito en
a i Medel estan ard cleclé on
that internal ‘model, whether otter
‘da ot pehotenivg arlion Ud bok incl of
fg alfecks ave «daw: Ok a _wwit of actions,
eal. : 3 deside. on_the basis of shat action
2 < and” decide. xt move’ and exe e
Akon + M ) mating 2. bowed on rest
fe 2 si iG p-eaced ; ced
: Pa cd ee models octai:-the fon of Ye
AOWVed 2 ndle|
i)
ie ewal
k ie | goals: ( Raticuety ued +. inplennent good 2 dey
ae 2 They aves weFlex+ node)
b> onc ako they see how ee
‘ Losi Ne Der fneedd) avo
“4Wot oltivate opel
Eg world ag tive now. Ti
eS
ws eating ction
ie k
Uti lity =z How Pigs wt
He
e
z be then > When we sc ABN aboutt itt ff
- Sat ef. we need.’ Computetional.
CAOVICES) » U ey ae D bts &
asl Ax * corsomedl performance &
0 Ge dene & Fioblen gererctte: > -Uility calewlate >
Gaels eruttnomenitad
- Xj ¥
sting, Sj Lenining “node part
ing, oF Coelho Lptoved _ perfovmarice standataae
> fexforn Dieiy + ieee (uve 2 caloolate -
wba Sov ae not) + :
2 _(Critins of Endtorment) 2» Qn _+hese
basis. clecidle: next wove -
5 => (ML agents work on aie rep
s “OF : Agen Ss ptt }
, x Carnot clecide (udhab actions re be exiting)
" al N otluation a sk i
ae any Qction seein Be petal i
J ine. of. bey i Oecie xno mc SONS :
_ hold. . “= pettatared) + \'
lip totgek
t AGE + Gupef
(Auiticiod ae 1 ral
fig to design wantindioval: agents ..
[hey ‘Ger: \eo ning agents, in for
g et G 10h
ob CHO a0 sepaiiertys .
° ° eo !
Oquakie: generauly |v elligenk.s
N 2
mvtiple
“capabiihy 0 gefiox Pass
- NLP f (nab
“Knowkdge Re prosertottin:
to Comin cite tn Eng 1°.
fe fing Test
f
ae < Atisretal Intelligence, es 3
Less A ts. the technology thort — ewolbles ‘sf
4 a Computer 40 tink of Oct utcehy
i ia a “Noman - use ok nviran wanets SY
i ee ke
“x Intellig ence 2 ? a>)
: P 2 J =
I. A Tink f s.
| Dy
= Tor’ e
x vting Test: cig ienaalg)- =
| —— = isa rnethod 2p =
<— of whot # hear:
Reasoving :
Wanows —
Ayio BOI
Use: to stow. “fafa
Os 2) 40 -dfow>_rew _concheisfons.-
— to answer
ye Machine Leaking ~ Z
; Nt adapt new __cimumstps
Tot g oraveces 2 sat: pietes
aeitizaal debination of tatoll garce |
me > T+ avoids py soal fateyaction betuxer
homan intertngator — the computer
becavse:- poutenl eouaiioa_of a
pertcon i. ot trite nce
_ To pass"! Tatag nest, she computer
neds “ eae ae 5
i Computay. Nigon + ‘To percep “objects
_ Robotics: To mans late _obfeaks
= ots A_move = dost +
‘ Tadnlting Ratfenally o. este
yy 20ses Stmbolic.
logic. “xo Coptie the lows Of & H
as Saricds that Con be ria olcked x Thiaving Huma . = fogntte_eabiitg
apprecicn J
3- —woyss oo out fra ae
| ins =
PFAANa TA
bein >
4_x Tottosee ction : (hy 5 i own 4} }.-
- Plgy chological. etiments Oot :
Be in =D aain
1 Obewh fn“ act
eel Neel” ah Etey Senor levslecec]
GRS ¢ beret)” Bisblemn Sher) -
>>the ale conceired tpysarel compaiving ©
the ore of Leaspring Sops LOK
= toes =of homon Sublect'S ” SAving
= the: game _ problems =
x Acting. Ratioray
>~ (Agent: comes Part lakn
= Ageten 0 oe)
E Rated A Agent «
; Bishe. Sree - 1S Se ie: an
ate SO = 0S 10 Othiop: <-thes-—bect
Wt @ne._of, when: thee te Onkefainit
he Best. 4 Br ‘
evil tft
is_to_tea.son logical to the Conclusion =
_¥ Cofect infevance ts not all of
Yotfona Lt:
> stiong Adee =
Con __wnachines have emotions? -
2s Ta case a8 biochem ee
yma ox vn ot Ihave +
eSAs— machines donat bow Biochem yeaction:
_* “ES On saadwire echoes AS
intedligorrtly as_a hunan_ beings
then tt iS. as intebl?gent as 2
Homan being’ ._(CTosing comesfon)-
Iv ase of Se NSAHON 3. Mac!
have. emotions )
$0: they don on
otters
~ Q). Why ee fons are
ANB tne ft helps us 48
with Dihera, trod —plal an_im
tole 2 y -
60. soning 5 =phoceSS 4
rode) ‘he man 72.0 n=
“‘eoctions ay Spock athe
pelavrouts QQ) Con machires ust! tove"- alex
Moore's Law: Ee HasiiBy of
—AYensistots that con _be-~ ft -
bY -O Computer chip wrt! deuble
Neh 18 months! Speer! of.
= Cory) puter creases ~2ad bern at
Dub le.
i
t
-Q): re ep = < whl =
= .Gbsorb = factor “a6 Cont patecl to. FF
A Our foekothess.? ete r
Sheed |
= abil ly of .an iets ag
to understand of leat. Aang,
Mtelle of tosk: : is the Spwserctation pf geretaltzed
. C Cognitive Abilittes in goPlsore
go that, Laced with an _nfomifor
fos\t athe AT _sastem could fad _|
ese SOON ai aescay Ez
x “Real AT.
Au Serious = Sciences
exa) — puipose AD
hard -
Spectr ad- =peipese AL ig mow doable
_(nba- teviol) = Eensible .
( Chess ; poker, voice tecpgrition »
= automate ~ Hanslation)-
x Intelligence -
handfot mater can
0 °
eAiCh. sf man pulate
e cormpliag exPetiorve y’. E : s
Ps Cay -to a ot eae =
OS Cheats, vy 4
intel Sersnal)
SVs +
anid _Yeason - =
Bote Absitity = 40 == understend - nid 0
ib ——aith. people", objects =2! Stmbolc —
p> Te. OFT te =Ock ee
“Vani slack won oll Zf
= a i
Wek arth nvivorment =
: abit
Comovter ° eile.
Sth ian
ae is aSS0cf ated ~ wit,
AeveSoping = Sdstem _ evolowed ea,
ae piocess ohare cferietica
ob 3 WinitanSasch anit
m
=¥ObOt -to itp... - ’ason ol tscover _meanin
generalize oF lean fem past
experienves »
Training Hbmanly AT:
3" The exciting new efFex+ 40 make
=computere think « machines ith mind
in_the - foll_ 3 tera) ence".
Bae bein Beet Havaplanch (485
fhe Automation oP acti ited] thot
we assectet with hyman- thinking »
acivites — gich- 0S _decish n-mra Ving »
_Pabbern solving, beoining 22”
Colman s (FFE
“gcfence +
“ - The_brosa_-as an
information _P teressing “machine »
- Rept donk theattes, Of bo
—bYoin wet S~
The Latter. tio methodologies “av
domains of cognitive eciere and
wanitive nev vosience-- :
: 2 ~Agia Homanly* e
"Ne . rt 20 3ng- Mackanes
ae “pean Gyactions — that vive inte LOgqorce n_ Ger forme
by. cope 2”
: HG, 10 hao ) 3
fie = ob ha a
Zz = (
ee Cree people Chaat eee “
Ase note 3 L
== Some intedlZeat. eae ray
AOL bes herman
= Homan Dsekinis a ss easy
to fool.
\
xe
rds upon.
' eas ee fale Chowiol_s) McDermott
(85) *
sh of Computations
tt possi ble ©
eason A o_act
a UNinston: Ay)
he D
FoF e
xno ods oo
ah dedocis
aopioach
oppy ee eh
Solve A i
eb hu
"i gnqularty.
TA G q
s The tom acca
rs oy field _o eoes$co) Phy
Ps > Denot@a. 2 Pot 1 Space. wlete.
Py —gasttational Fodd_hecannes iabeatte-
a of block de (we
4 wage Sind — a “singularity -
Pye Techindegy _si tid 3 pate
a2 ingelanty + =,
A 8 Sstoe ein
z
uy - A a itself te coipa heron _‘iwtellige nce!
(2 —chrangieg reas Ue OS we Know tt
jb x Computer _- fos psithout [teri ty Surpassing,
1 Homan fwtelig ence. (espaciatts vacontrolledy
VY esult— wills be Catastiophic ) -
Ve Hoon vind must merge
