machine(learning(|(stanford( design(and(analysis(of...

12
Ethan Solomon [email protected] MACHINE LEARNING | STANFORD DESIGN AND ANALYSIS OF ALGORITHMS | STANFORD PRINCIPLES OF MICROECONOMICS | UPENN COMPUTER SCIENCE & PROGRAMMING| MITx

Upload: others

Post on 01-Apr-2021

1 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: MACHINE(LEARNING(|(STANFORD( DESIGN(AND(ANALYSIS(OF ...linc.mit.edu/linc2013/presentations/LINC2013Solomon.pdf · RESISTANTVIRUS CLASS In order to model this effect, we introduce

Ethan  Solomon  [email protected]  

MACHINE  LEARNING  |  STANFORD  

DESIGN  AND  ANALYSIS  OF  ALGORITHMS  |  STANFORD  

PRINCIPLES  OF  MICROECONOMICS  |  UPENN  

COMPUTER  SCIENCE  &  PROGRAMMING|  MITx  

Page 2: MACHINE(LEARNING(|(STANFORD( DESIGN(AND(ANALYSIS(OF ...linc.mit.edu/linc2013/presentations/LINC2013Solomon.pdf · RESISTANTVIRUS CLASS In order to model this effect, we introduce

What  do  MOOCs  do  well?  

Page 3: MACHINE(LEARNING(|(STANFORD( DESIGN(AND(ANALYSIS(OF ...linc.mit.edu/linc2013/presentations/LINC2013Solomon.pdf · RESISTANTVIRUS CLASS In order to model this effect, we introduce

Lesson  1:  try,  try  again  

You  have  used  4  of  30  submissions  

(then  try  some  more)  

Page 4: MACHINE(LEARNING(|(STANFORD( DESIGN(AND(ANALYSIS(OF ...linc.mit.edu/linc2013/presentations/LINC2013Solomon.pdf · RESISTANTVIRUS CLASS In order to model this effect, we introduce

little  green  checkmarks  can  be  surprisingly  addicting  

Lesson  2:  

Page 5: MACHINE(LEARNING(|(STANFORD( DESIGN(AND(ANALYSIS(OF ...linc.mit.edu/linc2013/presentations/LINC2013Solomon.pdf · RESISTANTVIRUS CLASS In order to model this effect, we introduce

Lesson  3:  streamlined  is  better  MOOCs  are  more  efHicient  information  delivery  systems  

Page 6: MACHINE(LEARNING(|(STANFORD( DESIGN(AND(ANALYSIS(OF ...linc.mit.edu/linc2013/presentations/LINC2013Solomon.pdf · RESISTANTVIRUS CLASS In order to model this effect, we introduce

Lesson  4:  go  at  your  own  pace  (sort  of)  

Page 7: MACHINE(LEARNING(|(STANFORD( DESIGN(AND(ANALYSIS(OF ...linc.mit.edu/linc2013/presentations/LINC2013Solomon.pdf · RESISTANTVIRUS CLASS In order to model this effect, we introduce

Where  is  there  room  for  improvement?  

Page 8: MACHINE(LEARNING(|(STANFORD( DESIGN(AND(ANALYSIS(OF ...linc.mit.edu/linc2013/presentations/LINC2013Solomon.pdf · RESISTANTVIRUS CLASS In order to model this effect, we introduce

Feedback  is  deprived  with  autograders,  feedback  is  mostly  right-­‐or-­‐wrong.  

Page 9: MACHINE(LEARNING(|(STANFORD( DESIGN(AND(ANALYSIS(OF ...linc.mit.edu/linc2013/presentations/LINC2013Solomon.pdf · RESISTANTVIRUS CLASS In order to model this effect, we introduce

Little  creativity  in  lectures  in-­‐video  questions  are  nice,  but  

little  else  is  different  from    brick-­‐and-­‐mortar  lectures  

Page 10: MACHINE(LEARNING(|(STANFORD( DESIGN(AND(ANALYSIS(OF ...linc.mit.edu/linc2013/presentations/LINC2013Solomon.pdf · RESISTANTVIRUS CLASS In order to model this effect, we introduce

Course  organization  some  course  teams  may  initially  struggle  to  meet  the  demands  of  a  large  audience  or  their  own  self-­‐imposed  deadlines.  

Page 11: MACHINE(LEARNING(|(STANFORD( DESIGN(AND(ANALYSIS(OF ...linc.mit.edu/linc2013/presentations/LINC2013Solomon.pdf · RESISTANTVIRUS CLASS In order to model this effect, we introduce

21st  century  learning  

20th  century  campus  

Page 12: MACHINE(LEARNING(|(STANFORD( DESIGN(AND(ANALYSIS(OF ...linc.mit.edu/linc2013/presentations/LINC2013Solomon.pdf · RESISTANTVIRUS CLASS In order to model this effect, we introduce