Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
Cloud  Academy  &  AWS:  
how  we  use  Amazon  Web  Services  
for  machine  learning  and  data  collec:on
cloudacademy....
About  us
Alex  Casalboni Roberto  Turrin Luca  Baroffio
Sr.  SoCware  Engineer Sr.  Data  Scien:st  (PhD)   Data  Scien:st ...
What  is  Machine  Learning  (ML)?
Back  to  1959  (A.  Samuel)
Decision  problems  that    
can  be  modeled  from  data
...
Machine  Learning  pipeline
Training Predic1on
batch real-­‐:me
Feature  
extrac1on
batch
data informaGon
features ML  mod...
?
Machine  Learning  taxonomy
Supervised    
Learning
Unsupervised    
Learning
clda.co/webinar-ML
?
Machine  Learning  taxonomy
classifica3on
regression
170

cm
Supervised    
Learning
Unsupervised    
Learning
clda.co/we...
Machine  Learning  taxonomy
Supervised    
Learning
Unsupervised    
Learning
clda.co/webinar-ML
Machine  Learning  taxonomy
clustering
rule  extrac3on
group A group B
A, B C
Supervised    
Learning
Unsupervised    
Lea...
What  problems  can  ML  solve  for  you?
Supervised    
Learning
Unsupervised    
Learning
classifica'on
regression
cluste...
What  problems  can  ML  solve  for  you?
Supervised    
Learning
Unsupervised    
Learning
classifica'on
regression
cluste...
Learning
Data
Machine
Cloud
Big
Science
Information
Internet
Statistics
Technology
Python Future
Mining Social
Deep
IOT
Al...
Machine  learning  and  Big  data
“90%  of  the  data  in  the  world  today  has  been    
created  in  the  last  two  y...
Big  data  challenges
clda.co/webinar-ML
This  much  data  can’t  be  manually  inspected
Data-­‐driven  decisions
Distrib...
Why  is  deploying  ML  models  a  challenge?
clda.co/webinar-ML
Why  is  deploying  ML  models  a  challenge?
1.  Prototyping  !=  Produc=on-­‐ready
2.  We  need  Elas=city
4.  Avoid  la...
Where  is  the  lack  of  ownership?
clda.co/webinar-ML
!=
Data  Scien=st DevOps
Machine  Learning  
Data  mining  
Sta:s:...
Many  op:ons  and  tools  offered  by  AWS
ELB Auto  Scaling
Elas:c  
Beanstalk
Amazon  
ML
ECS
EMR LambdaEC2
API  
Gateway...
Serverless  compu:ng  to  the  rescue!
Transparent  scalability,  elas=city  and  availability
Developer-­‐friendly  maint...
How  is  “Serverless”  possible?
There is always a server somewhere,

you just don't have to worry about it :)
clda.co/web...
AWS  Lambda  +  Amazon  API  Gateway
+
AWS  
Lambda
API  
Gateway
RESTful  &  auth  layer
Global  CDN  and  caching  (Clou...
Quick  Example
clda.co/webinar-ML
clda.co/webinar-ML-example
clda.co/webinar-ML
clda.co/webinar-ML-lambda
AWS  Lambda  limita:ons
clda.co/webinar-ML
No  real-­‐=me  models  (only  pseudo  real-­‐=me)
Deployment  package  managem...
What  about  Amazon  Machine  Learning?
clda.co/webinar-ML
Amazon  
ML
One  of  the  first  MLaaS  solu=ons  (1  year  old)...
Key  Takeaways
clda.co/webinar-ML
Data-­‐driven  decision  and  user-­‐centered  ML  will  make  your  product  smarter
Ma...
Thank  you  for  acending  :)
cloudacademy.com
Q  &  A
Upcoming SlideShare
Loading in …5
×

Cloud Academy & AWS: how we use Amazon Web Services for machine learning and data collection

3,574 views

Published on

Speak with Alex Casalboni, Roberto Turrin and Luca Baroffio in our Engineering team at Cloud Academy, and learn how they use AWS to manage daily challenges and build a machine learning system.

Published in: Technology

Cloud Academy & AWS: how we use Amazon Web Services for machine learning and data collection

  1. 1. Cloud  Academy  &  AWS:   how  we  use  Amazon  Web  Services   for  machine  learning  and  data  collec:on cloudacademy.com 4/27/2016
  2. 2. About  us Alex  Casalboni Roberto  Turrin Luca  Baroffio Sr.  SoCware  Engineer Sr.  Data  Scien:st  (PhD)  Data  Scien:st  (PhD) @alex_casalboni @robytur @lucabaroffio clda.co/webinar-ML
  3. 3. What  is  Machine  Learning  (ML)? Back  to  1959  (A.  Samuel) Decision  problems  that     can  be  modeled  from  data clda.co/webinar-ML
  4. 4. Machine  Learning  pipeline Training Predic1on batch real-­‐:me Feature   extrac1on batch data informaGon features ML  models clda.co/webinar-ML
  5. 5. ? Machine  Learning  taxonomy Supervised     Learning Unsupervised     Learning clda.co/webinar-ML
  6. 6. ? Machine  Learning  taxonomy classifica3on regression 170
 cm Supervised     Learning Unsupervised     Learning clda.co/webinar-ML
  7. 7. Machine  Learning  taxonomy Supervised     Learning Unsupervised     Learning clda.co/webinar-ML
  8. 8. Machine  Learning  taxonomy clustering rule  extrac3on group A group B A, B C Supervised     Learning Unsupervised     Learning clda.co/webinar-ML
  9. 9. What  problems  can  ML  solve  for  you? Supervised     Learning Unsupervised     Learning classifica'on regression clustering rule  extrac'on ? 170
 cm gro gro A, B C clda.co/webinar-ML
  10. 10. What  problems  can  ML  solve  for  you? Supervised     Learning Unsupervised     Learning classifica'on regression clustering rule  extrac'on ? fraud  detecGon 170
 cm gro gro A, B C price  of  a  stock  over  Gme purchase  likelihood user  segmentaGon clda.co/webinar-ML
  11. 11. Learning Data Machine Cloud Big Science Information Internet Statistics Technology Python Future Mining Social Deep IOT Algorithms Management Storage Petabytes Parallel Network Privacy Million NoSQL PaaS SQL Database Exabytes Billion Dataset Hadoop R clda.co/webinar-ML
  12. 12. Machine  learning  and  Big  data “90%  of  the  data  in  the  world  today  has  been     created  in  the  last  two  years  alone”  -­‐  IBM “300+  hours  worth  of  video  content  is  being     uploaded  to  the  site  every  minute”  -­‐  Youtube clda.co/webinar-ML
  13. 13. Big  data  challenges clda.co/webinar-ML This  much  data  can’t  be  manually  inspected Data-­‐driven  decisions Distributed/parallel  compu=ng The  curse  of  dimensionality
  14. 14. Why  is  deploying  ML  models  a  challenge? clda.co/webinar-ML
  15. 15. Why  is  deploying  ML  models  a  challenge? 1.  Prototyping  !=  Produc=on-­‐ready 2.  We  need  Elas=city 4.  Avoid  lack  of  ownership clda.co/webinar-ML 3.  Too  many  nice-­‐to-­‐have  features
  16. 16. Where  is  the  lack  of  ownership? clda.co/webinar-ML != Data  Scien=st DevOps Machine  Learning   Data  mining   Sta:s:cal  analysis System  administra:on   (Cloud)  Opera:ons   SoCware  engineering
  17. 17. Many  op:ons  and  tools  offered  by  AWS ELB Auto  Scaling Elas:c   Beanstalk Amazon   ML ECS EMR LambdaEC2 API   Gateway clda.co/webinar-ML
  18. 18. Serverless  compu:ng  to  the  rescue! Transparent  scalability,  elas=city  and  availability Developer-­‐friendly  maintenance  (versioning  +  aliases) AWS   Lambda Event-­‐driven  approach  &  never  pay  for  idle 1  func=on  =  1  model clda.co/webinar-ML A/B  tes=ng  via  composi=on
  19. 19. How  is  “Serverless”  possible? There is always a server somewhere,
 you just don't have to worry about it :) clda.co/webinar-ML
  20. 20. AWS  Lambda  +  Amazon  API  Gateway + AWS   Lambda API   Gateway RESTful  &  auth  layer Global  CDN  and  caching  (CloudFront) Staging  &  versioning  &  mocking API  Decoupling clda.co/webinar-ML
  21. 21. Quick  Example clda.co/webinar-ML clda.co/webinar-ML-example
  22. 22. clda.co/webinar-ML clda.co/webinar-ML-lambda
  23. 23. AWS  Lambda  limita:ons clda.co/webinar-ML No  real-­‐=me  models  (only  pseudo  real-­‐=me) Deployment  package  management:  size  limit  and  OS  libraries Not  suitable  for  model  training  yet  (5  min  max  execu=on  =me)AWS   Lambda
  24. 24. What  about  Amazon  Machine  Learning? clda.co/webinar-ML Amazon   ML One  of  the  first  MLaaS  solu=ons  (1  year  old) Great  service  for  classifica=on  and  regression Only  linear  models  (linear  &  logis=c  regression  +  SGD) No  support  for  advanced  scenarios  yet     (collabora=ve  recommenda=on,  mul=media,  online  learning,  etc.)
  25. 25. Key  Takeaways clda.co/webinar-ML Data-­‐driven  decision  and  user-­‐centered  ML  will  make  your  product  smarter Maximize  ownership  by  removing  obstacles  btw  prototype  and  produc=on Eliminate  tradeoffs  btw  high-­‐scalability  and  nice-­‐to-­‐have  features Go  Serverless  and  stop  worrying  about  Ops MLaaS  makes  your  life  even  simpler,  unless  you  need  more  control
  26. 26. Thank  you  for  acending  :) cloudacademy.com Q  &  A

×