8/10/2019 A Seminar Report on Machine Learing
1/30
A Seminar Report on
MACHINE TRANSLATION
In the partial fulllment for the degree of B.Tech.
Seminar (8CS9)
JIET School of Engineering & Technology for Girls
epartment of Comp!ter Science & Engineering
"#$%'#$%
G!i*e* +y, '
S!-mitte* +y, ' Prof. Kamna Agarwal
Ms Meenakshi Soni Asst. Professor
IV ear !VIII Semester"
8/10/2019 A Seminar Report on Machine Learing
2/30
.MACHINE LEARNING / Seminar Report
ACKNOWLEDGEMENT
It is a matter of great pleasure for me to submit this report on MACHINE LEARNING, as a part
of curriculum for awar of !ACHEL"R#S IN $ECHN"L"G% &CSE' egree of Ra(asthan
$echnical )ni*ersit+, ota &Ra(asthan'-
At this moment of accomplishment, I am presenting m+ wor. with great prie an pleasure, I
woul li.e to e/press m+ sincere gratitue to all those who helpe me in the successful
completion of m+ *enture- I woul li.e to than. our PROF.KAMNA AGARWAL for helping
me in the successful accomplishment of m+ stu+ an for her timel+ an *aluable suggestions-
His constructi*e criticism has contribute immensel+ to the e*olution of m+ ieas on the sub(ect-
I am e/ceeingl+ grateful to m+ Head of Department PROF. MAMTA GARG an other
facult+ members for their inspiration an encouragement- I woul also li.e to than. m+ parents
an friens for their o*er whelming an whole hearte encouragement an support without
which this woul not ha*e been successful-
MEENAKSHI SONI
8/10/2019 A Seminar Report on Machine Learing
3/30
.MACHINE LEARNING / Seminar Report
JIET SCHOOL OF ENGINEERING TECHNOLOG! FOR GIRLS" JODHP#R
DEPARTMENT OF COMP#TER SCIENCE ENGINEERING
CERTIFICATE
$his is to certif+ that the report entitle $MACHINE LEARNING% has been carrie out
b+ MEENAKSHI SONI&nderm+ guiance in partial fulfillment of the egree of !achelor of
$echnolog+ in COMP#TER SCIENCE ENGINEERING of Ra(asthan $echnical
)ni*ersit+, ota uring the acaemic +ear012030124-
&5rof- ammna Agarwal'
Le't&rer E(am)ner
&5rof- Mamta Garg'
Head of Department
CSE
8/10/2019 A Seminar Report on Machine Learing
4/30
.MACHINE LEARNING / Seminar Report
A*STRACT
5resent a+ computer applications re6uire the representation of huge amount of comple/
.nowlege an ata in programs an thus re6uire tremenous amount of wor.- "ur abilit+ to
coe the computers falls short of the eman for applications- If the computers are enowe with
the learning abilit+, then our buren of coing the machine is ease &or at least reuce'- $his is
particularl+ true for e*eloping e/pert s+stems where the 7bottle3nec.7 is to e/tract the e/pert#s
.nowlege an fee the .nowlege to computers- $he present a+ computer programs in general
&with the e/ception of some Machine Learning programs' cannot correct their own errors or
impro*e from past mista.es, or learn to perform a new tas. b+ analog+ to a pre*iousl+ seen tas.-
In contrast, human beings are capable of all the abo*e- Machine Learning will prouce smarter
computers capable of all the abo*e intelligent beha*ior-
$he area of Machine Learning eals with the esign of programs that can learn rules from
ata, aapt to changes, an impro*e performance with e/perience- In aition to being one of the
initial reams of Computer Science, Machine Learning has become crucial as computers are
e/pecte to sol*e increasingl+ comple/ problems an become more integrate into our ail+
li*es- $his is a har problem, since ma.ing a machine learn from its computational tas.s re6uires
wor. at se*eral le*els, an comple/ities an ambiguities arise at each of those le*els-
So, here we stu+ how the Machine learning ta.e place, what are the methos, remeies
associate, applications, present an future status of machine learning-
8/10/2019 A Seminar Report on Machine Learing
5/30
.MACHINE LEARNING / Seminar Report
In*e0
ACN"8LE9GEMEN$
CER$I:ICA$E
A!S$RAC$$hapter % Introduction to Machine &earning '
%.% () MA$)I*+ &+A,*I*-
$hapter # &earning means /
#.% T)+ A,$)IT+$T0,+ 12 A &+A,*I*- A-+*T
$hapter 3 )istor4 of Machine leaning %#
3.% The *eural Modeling !Self 1rgani5ed S4stem"
3.# The S4m6olic $oncept Ac7uisition Paradigm3.3 The Modern Knowledge8Intensi9e Paradigm
$hapter : (ellsprings of Machine &earning %:
:.% Statistics
:.# Brain Models
:.3 Adapti9e $ontrol Theor4
:.: Ps4chological Models
:.; Articial Intelligence
:.' +9olutionar4 Models
$hapter ; Machine &earning 19er9iew %'
;.% The Aim of Machine &earning
;.# Machine &earning as a Science
$hapter ' $lassication of Machine &earning %amples of Machine &earning Pro6lems
$hapter < 2uture ?irections #' 9isco*er+ of new facts an theories through obser*ation an e/periment- :or
e/ample, the isco*er+ of ph+sics an chemistr+ laws-
$he general effect of learning in a s+stem is the impro*ement of the s+stem#s capabilit+
to sol*e problems- It is har to imagine a s+stem capable of learning cannot impro*e its problem3
sol*ing performance- A s+stem with learning capabilit+ shoul be able to o self3changing in
orer to perform better in its future problem3sol*ing-
8/10/2019 A Seminar Report on Machine Learing
10/30
Learning ? Impro*ing performance 5 at tas. $ b+
ac6uiring .nowlege using self3changing algorithm
A through e/perience E in an en*ironment for
tas. $-
.MACHINE LEARNING / Seminar Report
8e also note that earn)n0 'annot ta4e pa'e )n )/oat)on< 8e t+picall+ learn something
&.nowlege ' to perform some tas.s &$', through some e/perience E, an whether we ha*e
learne well or not will be (uge b+ some performance criteria 5 at the tas. $- :or e/ample, as
$om Mitchell put it in his ML boo., for the 7chec.ers learning problem7, the tas. $ is to pla+ the
game of chec.ers, the performance criteria 5 coul be the percentage of games won against
opponents, an the e/perience E coul be in the form pla+ing practice games with a teacher &or
self'- :or learning to ta.e place, we o nee a learning algorithm A for self3changing, which
allows the learner to get e/perience E in the tas. $, an ac6uire .nowlege &thus change the
learner#s .nowlege set' to impro*e the learner#s performance at tas. $-
$here are *arious forms of
impro*ement of a
s+stem#s problem3sol*ing abilit+
#uture Directions
Research in Machine Learning $heor+ is a combination of attac.ing establishe
funamental 6uestions, an e*eloping new framewor.s for moeling the nees of new machine
learning applications- 8hile it is impossible to .now where the ne/t brea.throughs will come, a
few topics one can e/pect the future to hol inclue.1 Conc)usions
Machine Learning $heor+ is both a funamental theor+ with man+ basic an compelling
founational 6uestions, an a topic of practical importance that helps to a*ance the state of the
art in software b+ pro*iing mathematical framewor.s for esigning new machine learning
algorithms- It is an e/citing time for the fiel, as connections to man+ other areas are being
isco*ere an e/plore, an as new machine learning applications bring new 6uestions to be
moele an stuie- It is safe to sa+ that the potential of Machine Learning an its theor+ lie
be+on the frontiers of our imagination-
8/10/2019 A Seminar Report on Machine Learing
30/30
.MACHINE LEARNING / Seminar Report
REFERENCES
@ Alpa+in, E- &011>'-!ntroduction to achine Learning. Massachusetts, )SA< MI$
5ress-
@ http