!"#$%&'()*(+,%(-.%$/,0(1,"2)$3'$/(4)(4.,(5627)"')$()8(

+4#3,$4(9$4,*,"4('$(-)&2#4,*(+:',$:,(

53(;%<)=">%(?'77(@(A,7'$3%(B%4,"(-.%'*('$(

-)&2#4,*(+:',$:,(@(5$/'$,,*'$/(C$'D,*"'4E()8(F%".'$/4)$(

5*':(1)G,*4"(H*)8,"")*()8(-)&2#4,*(+:',$:,(%$3(

?%""(C$'D,*"'4E(I,77)=('$(C$3,*/*%3#%4,(53#:%J)$(+4%$8)*3(C$'D,*"'4E(

(

K-F9!(LM4.(N$$'D,*"%*E(+#&&'4(A%E(OMLP(

0  

100  

200  

300  

400  

500  

600  

700  

800   Harvard  

CS50  

Introductory  course  enrollments  are  exploding  

0  100  200  300  400  500  600  700  800  900  1000   University  of  Pennsylvania  

total  

CIS  110  

CIS  120  

0  

200  

400  

600  

800  

1000  

1200   MIT  

total  

6.01  

6.00  

0  

100  

200  

300  

400  

500  

600  

700  

800   Harvard  

CS50  

Not  just  at  Harvard  

0  

500  

1000  

1500  

2000  

2500   Stanford  

total  

CS  106A  

CS  105  

CS  101  

Not  just  at  elite  private  insFtuFons  

0  

500  

1000  

1500  

2000  

2500  

3000   University  of  Washington  

CSE  142  

0  

200  

400  

600  

800  

1000  

1200  

1400   University  of  Michigan  

EECS  280  

These  enrollments  are  blowing  past  previous  highs  

0  

500  

1000  

1500  

2000  

2500  

3000   University  of  Washington  

CSE  142  

CSE  143  

Increasing  proporFons  of  students  are  taking  second  courses  

20%  

30%  

40%  

50%  

60%  

70%   University  of  Washington  

CSE  142-­‐>143  %  

In  at  least  some  cases,  female  parFcipaFon  is  increasing  

20%  

25%  

30%  

35%  

40%   University  of  Washington  

CSE  142  %  female  

0  

50  

100  

150  

200  

250  

300  

350   Harvard  

Majors  

0  

200  

400  

600  

800  

1000  

1200   MIT  

Total  

6.3  

6.2  

0  

100  

200  

300  

400  

500  

600   University  of  Pennsylvania  

Majors  

Demand  for  the  major  is  increasing  

0  

100  

200  

300  

400  

500  

600  

700  

800   Stanford  

Majors  

Dot-com peak: 3,237 Dot-bust trough: 1,732 Most recent year: 6,174 Target enrollment: 135 (+ ~40 upper-division transfers)

!"

#!!!"

$!!!"

%!!!"

&!!!"

'!!!"

(!!!"

)!!!"

#**'

"#*

*("

#**)

"#*

*+"

#***

"$!

!!"

$!!#

"$!

!$"

$!!%

"$!

!&"

$!!'

"$!

!("

$!!)

"$!

!+"

$!!*

"$!

#!"

$!##

"$!

#$"

$!#%

"$!

#&"

!"#$%&'%()%**+$(,$'-%#.'/0(

,-,"./012345"67789:45;1"

,-,"./012345"-8411"<4/=0;",9>0"

A1.O9:B10"279:.2"L10"3<<70O2=D=6=1."4130676"=6"=.407:6=.>Q":861"

!"

#!"

$!!"

$#!"

%!!"

%#!"

&!!"

'("$)$"*"+

,-./012"(

345-64"

'("$78"*"9

4-.":15-.;/

<-4"

'("$=$"*">

/<?@1-"A-

/.1@12"

'("$7$"*"B

@4C/D@

E/0F1"

'("$!8"*"G

/5/"(<

@-1<-"

!"#$"#%&'(()#*+,$,-,./&0.1#-)-&

H1.FDD6-15"

!"

#!"

$!"

%!"

&!"

'!!"

'#!"

#!'!(''" #!''('#" #!'#(')" #!')('$"

!"#$%&'()**'+,--%.%'+/'01'

*+,(-./+01"

+8:66"6=T76":07">1=.>"/I013>I"/I7"011L"

!"

#!"

$!!"

$#!"

%!!"

%#!"

&!!"

'("&)$!"*"'+,-./01"21345/03/.10"

'("))$!"*"6-01789:"(;</0,<"

'("#$#!"*"(+=>710"?9:5900159:"

!"#$%&&'())%#'*+,+-+"$'!".#-%-'

?91+@@,09/"

+8:66"6=T76":07">1=.>"/I013>I"/I7"011L"

!"

#!"

$!!"

$#!"

%!!"

%#!"

&!!"

'("&)$!"*"'+,-./01"21345/03/.10"

'("))$!"*"6-01789:"(;</0,<"

'("#$#!"*"(+=>710"?9:5900159:"

!"#$%&&'())%#'*+,+-+"$'!".#-%-'

?91+@@,09/"

760 students!

CS229!Machine Learning!Autumn 2013!

CS229Machine LearningAutumn 2013

Why is this man smiling?

U7V07":88"L:9=8=:0"K=/I"454876"=."279:.2"

!"

#!$!!!"

%!$!!!"

&!$!!!"

'!$!!!"

(!$!!!"

)!$!!!"

*!$!!!"

#+))

"#+

)*"

#+),

"#+

)+"

#+*!

"#+

*#"

#+*%

"#+

*&"

#+*'

"#+

*("

#+*)

"#+

**"

#+*,

"#+

*+"

#+,!

"#+

,#"

#+,%

"#+

,&"

#+,'

"#+

,("

#+,)

"#+

,*"

#+,,

"#+

,+"

#++!

"#+

+#"

#++%

"#+

+&"

#++'

"#+

+("

#++)

"#+

+*"

#++,

"#+

++"

%!!!

"%!

!#"

%!!%

"%!

!&"

%!!'

"%!

!("

%!!)

"%!

!*"

%!!,

"%!

!+"

%!#!

"

!"#$%&'()*+,'-+')./+0'1"(2)3'4(''2)5(/-&'6)

But  this  Fme,  something  feels  different  •  Possibly,  students  are  figuring  out  that  every  21st    century  ciFzen  

needs  to  have  facility  with  “computaFonal  thinking”  –  problem  analysis  and  decomposiFon  (stepwise  refinement),  abstracFon,  algorithmic  thinking,  algorithmic  expression,  stepwise  fault  isolaFon  (debugging),  modeling  –  Programming  is  the  hands-­‐on,  inquiry-­‐based  way  that  we  teach  

computaFonal  thinking  and  the  principles  of  computer  science  

•  Possibly,  students  are  figuring  out  that  fields  from  Anthropology  to  Zoology  are  becoming  informa(on  fields,  and  that  those  who  can  bend  the  power  of  the  computer  to  their  will  –  computaFonal  thinking,  but  also  computer  science  in  greater  depth  –  will  be  posiFoned  for  greater  success  than  those  who  can’t  

•! Y166=J85Q"6/327./6":07"Z>30=.>"13/"/I:/"419<3/70"64=7.47"=6".1/"S=8J70/"]"=/V6":."=./78874/3:885"7;4=F.>Q"I=>I85"407:FD7":.2"=./70:4FD7Q"[<1K70"/1"4I:.>7"/I7"K1082\"Z782"

•  Possibly,  students  are  figuring  out  that  all  of  the  STEM  jobs  are  in  computer  science  

Y166=J85Q"/I7"0747./"20:9:F4">01K/I"K=88"41.F.37"

!"

#!$!!!"

%!$!!!"

&!$!!!"

'!$!!!"

(!$!!!"

)!$!!!"

*!$!!!"

#+))

"#+

)*"

#+),

"#+

)+"

#+*!

"#+

*#"

#+*%

"#+

*&"

#+*'

"#+

*("

#+*)

"#+

**"

#+*,

"#+

*+"

#+,!

"#+

,#"

#+,%

"#+

,&"

#+,'

"#+

,("

#+,)

"#+

,*"

#+,,

"#+

,+"

#++!

"#+

+#"

#++%

"#+

+&"

#++'

"#+

+("

#++)

"#+

+*"

#++,

"#+

++"

%!!!

"%!

!#"

%!!%

"%!

!&"

%!!'

"%!

!("

%!!)

"%!

!*"

%!!,

"%!

!+"

%!#!

"

!"#$%&'()*+,'-+')./+0'1"(2)3'4(''2)5(/-&'6)

Y166=J85Q"/I7"0747./"20:9:F4">01K/I"K=88"41.F.37"

!"

#!$!!!"

%!$!!!"

&!$!!!"

'!$!!!"

(!$!!!"

)!$!!!"

*!$!!!"

#+))

"#+

)*"

#+),

"#+

)+"

#+*!

"#+

*#"

#+*%

"#+

*&"

#+*'

"#+

*("

#+*)

"#+

**"

#+*,

"#+

*+"

#+,!

"#+

,#"

#+,%

"#+

,&"

#+,'

"#+

,("

#+,)

"#+

,*"

#+,,

"#+

,+"

#++!

"#+

+#"

#++%

"#+

+&"

#++'

"#+

+("

#++)

"#+

+*"

#++,

"#+

++"

%!!!

"%!

!#"

%!!%

"%!

!&"

%!!'

"%!

!("

%!!)

"%!

!*"

%!!,

"%!

!+"

%!#!

"

!"#$%&'()*+,'-+')./+0'1"(2)3'4(''2)5(/-&'6)

How  are  we,  and  our  insFtuFons,  going  to  respond?  •  10%  of  Princeton’s  students  are  computer  science  majors  

–  Far  more  at  MIT  

•  10%  of  Princeton’s  faculty  are  unlikely  to  ever  be  in  computer  science!  –  Dieo,  proporFonately,  at  MIT  

•  And  then  there  is  –  Introductory  course  demand  ++  –  Upper-­‐division  non-­‐major  demand  ++  –  Graduate  non-­‐major  demand  ++  

Some  possibiliFes  …  •  Restrict  the  size  of  the  major  

–  ImplicaFons  for  diversity?  

•  Exclude  non-­‐majors  from  upper-­‐division  courses  •  Retreat  to  “the  core”  –  turn  over  many  of  our  courses  to  other  

departments  •  Have  enormous  class  sizes  and/or  enormous  teaching  loads  •  UFlize  vast  numbers  of  lecturers  •  Have  a  beer  while  the  students  use  Coursera  

That’s  what  we’ll  discuss  •  What  are  you  experiencing?  •  How  are  you  responding  today?  •  What  do  you  envision  in  the  short-­‐term  and  long-­‐term  future?  •  What  are  the  possible  implicaFons  of  various  direcFons  for  the  

parFcipaFon  of  currently  under-­‐represented  groups?  

Our  field  is  at  a  criFcal  juncture!  

http://lazowska.cs.washington.edu/NCWIT.pdf

Thanks  for  providing  data!  •  Colorado  School  of  Mines:  Tracy  Camp  •  Harvard  Univ.:  Greg  Morrisee,  Margo  Seltzer  •  Harvey  Mudd  College:  Ran  Libeskind-­‐Hadas  •  Johns  Hopkins  Univ.:  Greg  Hager  •  MIT:  Saman  Amarasinghe,  Anantha  Chandrakasan,  Eric  Grimson,  Laura  Moses,  Victoria  

Palay,  Daniela  Rus,  Jacob  White  •  Rochester  InsFtute  of  Technology:  Andrew  Sears  •  Stanford  Univ.:  Eric  Roberts,  Claire  Stager  •  Univ.  Michigan:  H.V.  Jagadish  •  Univ.  Pennsylvania:  Sue  Davidson  •  Univ.  Rochester:  Henry  Kautz  •  Univ.  Texas:    Tiffany  Grady,  J  Moore  •  Univ.  Utah:  Ross  Whitaker  •  Univ.  Washington:  Raven  Alexander,  Crystal  Eney,  Ed  Lazowska,  Jen  Pesicka  •  Wellesley  College:  Takis  Metaxas