SlideShare uma empresa Scribd logo
1 de 6
Baixar para ler offline
Deep	
  Learning	
  For	
  Prac//oners	
  	
  
Lecture	
  2:	
  Which	
  applica/ons	
  benefit	
  from	
  
deep	
  learning?	
  
Anantharaman	
  Narayana	
  Iyer	
  	
  
deeplearning.ananth@gmail.com	
  
17th	
  June	
  2014	
  
Note:	
  Notes	
  that	
  contain	
  code	
  examples	
  for	
  these	
  slides	
  and	
  
detailed	
  analysis	
  will	
  be	
  published	
  separately	
  later.	
  
Review	
  of	
  previous	
  lecture	
  
•  Deep	
  learning	
  as	
  a	
  major	
  machine	
  learning	
  discipline	
  
has	
  received	
  phenomenal	
  aNen/on	
  of	
  late	
  due	
  to:	
  
–  Breakthrough	
  results	
  reported	
  by	
  the	
  research	
  
community	
  	
  for	
  certain	
  class	
  of	
  applica/ons,	
  beNering	
  
the	
  current	
  state	
  of	
  the	
  art	
  
–  Substan/al	
  investments	
  by	
  technology	
  companies	
  
such	
  as:	
  Google,	
  Facebook,	
  MicrosoU,	
  IBM	
  	
  
•  While	
  there	
  is	
  no	
  single	
  unique	
  architecture,	
  deep	
  
networks	
  are	
  typically	
  built	
  using	
  some	
  variant	
  of	
  
Autoencoders	
  or	
  Restricted	
  Boltzmann	
  Machines	
  
with	
  key	
  characteris/cs	
  of:	
  
–  Deep	
  architecture:	
  Mul/ple	
  layers	
  performing	
  
complex,	
  nonlinear	
  computa/ons,	
  cascading	
  the	
  
layerwise	
  outputs.	
  
–  Automated	
  feature	
  extrac/on:	
  each	
  layer	
  produces	
  as	
  
its	
  output	
  an	
  abstracted	
  form	
  of	
  its	
  inputs	
  (e.g.	
  Edges	
  
from	
  raw	
  pixels).	
  One	
  may	
  add	
  a	
  classifier	
  layer	
  (e.g	
  
SVM)	
  on	
  top	
  of	
  the	
  abstracted	
  features	
  and	
  can	
  view	
  
the	
  classifica/on	
  as	
  being	
  done	
  on	
  the	
  most	
  abstract	
  
features	
  automa/cally	
  generated	
  by	
  the	
  system.	
  (An	
  
example	
  with	
  code	
  illustrated	
  in	
  the	
  next	
  lecture)	
  
Looking	
  through	
  the	
  prac//oner’s	
  prism	
  
•  To	
  address	
  real	
  world	
  problems,	
  
prac//oners	
  need	
  to	
  be	
  aware	
  of	
  where	
  
deep	
  learning	
  yields	
  best	
  results,	
  prac/cal	
  
considera/ons,	
  limita/ons	
  and	
  when	
  not	
  to	
  
use	
  it.	
  
	
  
•  This	
  requires	
  looking	
  at	
  the	
  research	
  results	
  
and	
  other	
  claims	
  from	
  a	
  prac/cal	
  
perspec/ve	
  and	
  stay	
  clear	
  of	
  common	
  
misconcep/ons.	
  
“If	
  all	
  you	
  have	
  is	
  a	
  hammer	
  everything	
  looks	
  as	
  a	
  nail”	
  
•  Deep	
  learning	
  has	
  proved	
  its	
  poten/al	
  in	
  some	
  applica/on	
  domains	
  (e.g.	
  
Computer	
  Vision,	
  Speech	
  recogni/on),	
  holds	
  early	
  promise	
  in	
  several	
  other	
  areas	
  
(e.g	
  Natural	
  Language	
  Processing)	
  but	
  this	
  is	
  not	
  a	
  universal	
  tool	
  to	
  provide	
  the	
  
best	
  result	
  for	
  “any”	
  AI	
  task.	
  
•  When	
  does	
  it	
  have	
  the	
  poten/al	
  to	
  perform	
  best?	
  
–  When	
  structure	
  of	
  the	
  problem	
  being	
  solved	
  naturally	
  maps	
  to	
  a	
  mul/	
  layer	
  
architecture	
  
•  If	
  the	
  problem	
  we	
  are	
  trying	
  to	
  solve	
  can	
  be	
  decomposed	
  in	
  to	
  processing	
  hierarchical	
  
abstract	
  features	
  and	
  these	
  features	
  are	
  derivable	
  from	
  the	
  input	
  data	
  through	
  a	
  set	
  of	
  
poten/ally	
  nonlinear	
  transforma/ons,	
  deep	
  learning	
  based	
  solu/on	
  might	
  be	
  effec/ve.	
  	
  
•  As	
  a	
  corollary,	
  problems	
  that	
  don’t	
  exhibit	
  a	
  mul/	
  layer	
  structure	
  may	
  not	
  see	
  much	
  
incremental	
  benefit	
  compared	
  to	
  tradi/onal	
  methods	
  
–  Data	
  availability	
  
•  While	
  tradi/onal	
  architectures	
  require	
  expert	
  designed	
  features,	
  deep	
  learning	
  systems	
  
automa/cally	
  learn	
  these	
  features,	
  given	
  the	
  raw	
  input.	
  
•  In	
  order	
  to	
  learn	
  the	
  features,	
  extensive,	
  unsupervised	
  pretraining	
  using	
  large	
  volumes	
  of	
  
data	
  is	
  oUen	
  required.	
  Hence	
  any	
  advanced	
  solu/on	
  based	
  on	
  deep	
  learning	
  is	
  likely	
  to	
  
require	
  availability	
  of	
  such	
  data.	
  
“More	
  data	
  or	
  beNer	
  models?”	
  
•  Data	
  Vs	
  Algorithm:	
  research	
  shows	
  that	
  
training	
  a	
  system	
  with	
  more	
  data,	
  the	
  
performance	
  asympto/cally	
  approaches	
  
same	
  levels	
  regardless	
  of	
  the	
  model.	
  
•  One	
  may	
  be	
  led	
  to	
  believe	
  that	
  shallow	
  
networks,	
  trained	
  with	
  huge	
  data	
  might	
  
equal	
  the	
  performance	
  of	
  deep	
  networks.	
  	
  
–  Unfortunately,	
  much	
  of	
  the	
  available	
  data	
  in	
  the	
  
web	
  is	
  unlabeled	
  and	
  without	
  an	
  effec/ve	
  
unsupervised	
  training	
  model,	
  the	
  data	
  is	
  not	
  
useful.	
  Deep	
  networks	
  with	
  unsupervised	
  
pretraining	
  phase,	
  can	
  leverage	
  the	
  data	
  beNer.	
  
•  Another	
  no/on	
  could	
  be	
  that	
  any	
  
algorithm	
  or	
  model	
  selec/on	
  for	
  a	
  deep	
  
network	
  is	
  good	
  enough	
  if	
  you	
  give	
  it	
  a	
  
huge	
  volume	
  of	
  data.	
  
–  Choosing	
  an	
  op/mal	
  algorithm	
  and	
  design	
  is	
  
very	
  cri/cal	
  as	
  deep	
  networks	
  are	
  resource	
  
heavy	
  due	
  to	
  mul/ple	
  layers	
  and	
  weights.	
  A	
  
good	
  intui/on	
  on	
  the	
  problem	
  structure	
  is	
  
important	
  to	
  make	
  right	
  choices	
  of	
  the	
  model.	
  
Automated	
  Feature	
  Learning	
  and	
  data	
  preprocessing	
  
Though	
  deep	
  learning	
  systems	
  extract	
  features	
  automa/cally,	
  the	
  task	
  of	
  data	
  
preprocessing	
  is	
  s/ll	
  non-­‐trivial.	
  
–  The	
  input	
  data	
  should	
  be	
  complete	
  enough	
  so	
  that	
  the	
  features	
  relevant	
  for	
  the	
  given	
  
problem	
  can	
  be	
  extracted.	
  
•  Consider	
  the	
  example	
  of	
  detec/ng	
  anomalies	
  in	
  the	
  opera/on	
  of	
  a	
  nuclear	
  reactor.	
  The	
  
input	
  to	
  be	
  given	
  to	
  a	
  deep	
  learning	
  system	
  should	
  include	
  signals	
  from	
  all	
  the	
  relevant	
  
sensors	
  and	
  missing	
  any	
  of	
  them	
  may	
  result	
  in	
  inadequate	
  performance	
  
–  The	
  op/mum	
  size	
  of	
  the	
  input	
  data	
  adequate	
  for	
  the	
  job	
  needs	
  to	
  be	
  determined.	
  
•  Suppose	
  we	
  need	
  to	
  perform	
  face	
  detec/on,	
  given	
  the	
  input	
  images.	
  What	
  should	
  be	
  the	
  
right	
  input	
  size?	
  Should	
  it	
  be	
  10	
  x	
  10	
  or	
  100	
  x	
  100	
  pixels?	
  High	
  dimensionality	
  increases	
  the	
  
model	
  parameters	
  substan/ally,	
  requiring	
  more	
  compute	
  resources.	
  
–  Input	
  vector	
  representa/on	
  must	
  be	
  determined	
  
•  E.g,	
  for	
  an	
  NLP	
  problem,	
  words	
  from	
  a	
  vocabulary	
  V	
  may	
  be	
  represented	
  in	
  “one-­‐hot”	
  form	
  
where	
  each	
  word	
  in	
  V	
  is	
  represented	
  by	
  a	
  posi/on.	
  Here,	
  the	
  number	
  of	
  features	
  for	
  a	
  given	
  
word	
  w	
  equals	
  the	
  size	
  of	
  	
  the	
  vocabulary	
  |V|	
  and	
  a	
  sentence	
  with	
  k	
  words	
  will	
  be	
  
represented	
  as	
  k	
  *	
  |V|	
  sized	
  input	
  vector.	
  When	
  the	
  size	
  of	
  vocabulary	
  becomes	
  large	
  (say	
  
over	
  10000	
  words),	
  this	
  representa/on	
  increases	
  the	
  dimensionality	
  substan/ally.	
  
–  For	
  many	
  problems,	
  data	
  cleaning	
  and	
  preprocessing	
  are	
  s/ll	
  required	
  
•  E.g.	
  For	
  many	
  NLP	
  problems,	
  beNer	
  performance	
  may	
  be	
  obtained	
  easier	
  through	
  some	
  
preprocessing	
  steps	
  (such	
  as	
  stopword	
  removal,	
  stemming	
  etc)	
  rather	
  than	
  lehng	
  the	
  deep	
  
learning	
  system	
  handle	
  the	
  data	
  in	
  its	
  raw	
  form.	
  

Mais conteúdo relacionado

Mais procurados

Natural Language Processing Advancements By Deep Learning: A Survey
Natural Language Processing Advancements By Deep Learning: A SurveyNatural Language Processing Advancements By Deep Learning: A Survey
Natural Language Processing Advancements By Deep Learning: A SurveyRimzim Thube
 
Introduction to Machine Learning
Introduction to Machine LearningIntroduction to Machine Learning
Introduction to Machine LearningShahar Cohen
 
MMDS 2014 Talk - Distributing ML Algorithms: from GPUs to the Cloud
MMDS 2014 Talk - Distributing ML Algorithms: from GPUs to the CloudMMDS 2014 Talk - Distributing ML Algorithms: from GPUs to the Cloud
MMDS 2014 Talk - Distributing ML Algorithms: from GPUs to the CloudXavier Amatriain
 
Generating Natural-Language Text with Neural Networks
Generating Natural-Language Text with Neural NetworksGenerating Natural-Language Text with Neural Networks
Generating Natural-Language Text with Neural NetworksJonathan Mugan
 
Deep Learning Enabled Question Answering System to Automate Corporate Helpdesk
Deep Learning Enabled Question Answering System to Automate Corporate HelpdeskDeep Learning Enabled Question Answering System to Automate Corporate Helpdesk
Deep Learning Enabled Question Answering System to Automate Corporate HelpdeskSaurabh Saxena
 
"An Introduction to Machine Learning and How to Teach Machines to See," a Pre...
"An Introduction to Machine Learning and How to Teach Machines to See," a Pre..."An Introduction to Machine Learning and How to Teach Machines to See," a Pre...
"An Introduction to Machine Learning and How to Teach Machines to See," a Pre...Edge AI and Vision Alliance
 
Machine Learning Lecture 2 Basics
Machine Learning Lecture 2 BasicsMachine Learning Lecture 2 Basics
Machine Learning Lecture 2 Basicsananth
 
Deep learning Introduction and Basics
Deep learning  Introduction and BasicsDeep learning  Introduction and Basics
Deep learning Introduction and BasicsNitin Mishra
 
Muhammad Usman Akhtar | Ph.D Scholar | Wuhan University | School of Co...
Muhammad Usman Akhtar  |  Ph.D Scholar  |  Wuhan  University  |  School of Co...Muhammad Usman Akhtar  |  Ph.D Scholar  |  Wuhan  University  |  School of Co...
Muhammad Usman Akhtar | Ph.D Scholar | Wuhan University | School of Co...Wuhan University
 
Machine Learning Unit 1 Semester 3 MSc IT Part 2 Mumbai University
Machine Learning Unit 1 Semester 3  MSc IT Part 2 Mumbai UniversityMachine Learning Unit 1 Semester 3  MSc IT Part 2 Mumbai University
Machine Learning Unit 1 Semester 3 MSc IT Part 2 Mumbai UniversityMadhav Mishra
 
L06 stemmer and edit distance
L06 stemmer and edit distanceL06 stemmer and edit distance
L06 stemmer and edit distanceananth
 
A tutorial on deep learning at icml 2013
A tutorial on deep learning at icml 2013A tutorial on deep learning at icml 2013
A tutorial on deep learning at icml 2013Philip Zheng
 
NLP Classifier Models & Metrics
NLP Classifier Models & MetricsNLP Classifier Models & Metrics
NLP Classifier Models & MetricsSanghamitra Deb
 
ODSC East: Effective Transfer Learning for NLP
ODSC East: Effective Transfer Learning for NLPODSC East: Effective Transfer Learning for NLP
ODSC East: Effective Transfer Learning for NLPindico data
 
Machine Learning Applications in NLP.ppt
Machine Learning Applications in NLP.pptMachine Learning Applications in NLP.ppt
Machine Learning Applications in NLP.pptbutest
 
Introduction to Machine Learning
Introduction to Machine LearningIntroduction to Machine Learning
Introduction to Machine LearningShimi Bandiel
 
Deep Learning Architectures for NLP (Hungarian NLP Meetup 2016-09-07)
Deep Learning Architectures for NLP (Hungarian NLP Meetup 2016-09-07)Deep Learning Architectures for NLP (Hungarian NLP Meetup 2016-09-07)
Deep Learning Architectures for NLP (Hungarian NLP Meetup 2016-09-07)Márton Miháltz
 
Understanding Basics of Machine Learning
Understanding Basics of Machine LearningUnderstanding Basics of Machine Learning
Understanding Basics of Machine LearningPranav Ainavolu
 
H transformer-1d paper review!!
H transformer-1d paper review!!H transformer-1d paper review!!
H transformer-1d paper review!!taeseon ryu
 

Mais procurados (20)

Natural Language Processing Advancements By Deep Learning: A Survey
Natural Language Processing Advancements By Deep Learning: A SurveyNatural Language Processing Advancements By Deep Learning: A Survey
Natural Language Processing Advancements By Deep Learning: A Survey
 
Introduction to Machine Learning
Introduction to Machine LearningIntroduction to Machine Learning
Introduction to Machine Learning
 
MMDS 2014 Talk - Distributing ML Algorithms: from GPUs to the Cloud
MMDS 2014 Talk - Distributing ML Algorithms: from GPUs to the CloudMMDS 2014 Talk - Distributing ML Algorithms: from GPUs to the Cloud
MMDS 2014 Talk - Distributing ML Algorithms: from GPUs to the Cloud
 
Generating Natural-Language Text with Neural Networks
Generating Natural-Language Text with Neural NetworksGenerating Natural-Language Text with Neural Networks
Generating Natural-Language Text with Neural Networks
 
Deep Learning Enabled Question Answering System to Automate Corporate Helpdesk
Deep Learning Enabled Question Answering System to Automate Corporate HelpdeskDeep Learning Enabled Question Answering System to Automate Corporate Helpdesk
Deep Learning Enabled Question Answering System to Automate Corporate Helpdesk
 
"An Introduction to Machine Learning and How to Teach Machines to See," a Pre...
"An Introduction to Machine Learning and How to Teach Machines to See," a Pre..."An Introduction to Machine Learning and How to Teach Machines to See," a Pre...
"An Introduction to Machine Learning and How to Teach Machines to See," a Pre...
 
Machine Learning Lecture 2 Basics
Machine Learning Lecture 2 BasicsMachine Learning Lecture 2 Basics
Machine Learning Lecture 2 Basics
 
Deep learning Introduction and Basics
Deep learning  Introduction and BasicsDeep learning  Introduction and Basics
Deep learning Introduction and Basics
 
Scene understanding
Scene understandingScene understanding
Scene understanding
 
Muhammad Usman Akhtar | Ph.D Scholar | Wuhan University | School of Co...
Muhammad Usman Akhtar  |  Ph.D Scholar  |  Wuhan  University  |  School of Co...Muhammad Usman Akhtar  |  Ph.D Scholar  |  Wuhan  University  |  School of Co...
Muhammad Usman Akhtar | Ph.D Scholar | Wuhan University | School of Co...
 
Machine Learning Unit 1 Semester 3 MSc IT Part 2 Mumbai University
Machine Learning Unit 1 Semester 3  MSc IT Part 2 Mumbai UniversityMachine Learning Unit 1 Semester 3  MSc IT Part 2 Mumbai University
Machine Learning Unit 1 Semester 3 MSc IT Part 2 Mumbai University
 
L06 stemmer and edit distance
L06 stemmer and edit distanceL06 stemmer and edit distance
L06 stemmer and edit distance
 
A tutorial on deep learning at icml 2013
A tutorial on deep learning at icml 2013A tutorial on deep learning at icml 2013
A tutorial on deep learning at icml 2013
 
NLP Classifier Models & Metrics
NLP Classifier Models & MetricsNLP Classifier Models & Metrics
NLP Classifier Models & Metrics
 
ODSC East: Effective Transfer Learning for NLP
ODSC East: Effective Transfer Learning for NLPODSC East: Effective Transfer Learning for NLP
ODSC East: Effective Transfer Learning for NLP
 
Machine Learning Applications in NLP.ppt
Machine Learning Applications in NLP.pptMachine Learning Applications in NLP.ppt
Machine Learning Applications in NLP.ppt
 
Introduction to Machine Learning
Introduction to Machine LearningIntroduction to Machine Learning
Introduction to Machine Learning
 
Deep Learning Architectures for NLP (Hungarian NLP Meetup 2016-09-07)
Deep Learning Architectures for NLP (Hungarian NLP Meetup 2016-09-07)Deep Learning Architectures for NLP (Hungarian NLP Meetup 2016-09-07)
Deep Learning Architectures for NLP (Hungarian NLP Meetup 2016-09-07)
 
Understanding Basics of Machine Learning
Understanding Basics of Machine LearningUnderstanding Basics of Machine Learning
Understanding Basics of Machine Learning
 
H transformer-1d paper review!!
H transformer-1d paper review!!H transformer-1d paper review!!
H transformer-1d paper review!!
 

Destaque

Natural Language Processing: L03 maths fornlp
Natural Language Processing: L03 maths fornlpNatural Language Processing: L03 maths fornlp
Natural Language Processing: L03 maths fornlpananth
 
Natural Language Processing: L02 words
Natural Language Processing: L02 wordsNatural Language Processing: L02 words
Natural Language Processing: L02 wordsananth
 
Overview of TensorFlow For Natural Language Processing
Overview of TensorFlow For Natural Language ProcessingOverview of TensorFlow For Natural Language Processing
Overview of TensorFlow For Natural Language Processingananth
 
Natural Language Processing: L01 introduction
Natural Language Processing: L01 introductionNatural Language Processing: L01 introduction
Natural Language Processing: L01 introductionananth
 
L05 language model_part2
L05 language model_part2L05 language model_part2
L05 language model_part2ananth
 
Convolutional Neural Networks: Part 1
Convolutional Neural Networks: Part 1Convolutional Neural Networks: Part 1
Convolutional Neural Networks: Part 1ananth
 
Word representation: SVD, LSA, Word2Vec
Word representation: SVD, LSA, Word2VecWord representation: SVD, LSA, Word2Vec
Word representation: SVD, LSA, Word2Vecananth
 
Deep Learning For Speech Recognition
Deep Learning For Speech RecognitionDeep Learning For Speech Recognition
Deep Learning For Speech Recognitionananth
 
Present simple tense
Present simple tensePresent simple tense
Present simple tenseptnphuong
 
Present simple tense
Present simple tensePresent simple tense
Present simple tenseptnphuong
 
Introduction and Starting ASP.NET MVC
Introduction and Starting ASP.NET MVCIntroduction and Starting ASP.NET MVC
Introduction and Starting ASP.NET MVCYogendra Tamang
 
Azure machine learning tech mela
Azure machine learning tech melaAzure machine learning tech mela
Azure machine learning tech melaYogendra Tamang
 
Operations Management Study in Textured Jersy Lanka Limited
Operations Management Study in Textured Jersy Lanka LimitedOperations Management Study in Textured Jersy Lanka Limited
Operations Management Study in Textured Jersy Lanka LimitedHansi Thenuwara
 
Machine learning and azure ml studio
Machine learning and azure ml studioMachine learning and azure ml studio
Machine learning and azure ml studioYogendra Tamang
 
L05 word representation
L05 word representationL05 word representation
L05 word representationananth
 
Machine learning and azure ml studio gabc
Machine learning and azure ml studio gabcMachine learning and azure ml studio gabc
Machine learning and azure ml studio gabcYogendra Tamang
 

Destaque (20)

Natural Language Processing: L03 maths fornlp
Natural Language Processing: L03 maths fornlpNatural Language Processing: L03 maths fornlp
Natural Language Processing: L03 maths fornlp
 
Natural Language Processing: L02 words
Natural Language Processing: L02 wordsNatural Language Processing: L02 words
Natural Language Processing: L02 words
 
Overview of TensorFlow For Natural Language Processing
Overview of TensorFlow For Natural Language ProcessingOverview of TensorFlow For Natural Language Processing
Overview of TensorFlow For Natural Language Processing
 
Natural Language Processing: L01 introduction
Natural Language Processing: L01 introductionNatural Language Processing: L01 introduction
Natural Language Processing: L01 introduction
 
L05 language model_part2
L05 language model_part2L05 language model_part2
L05 language model_part2
 
Convolutional Neural Networks: Part 1
Convolutional Neural Networks: Part 1Convolutional Neural Networks: Part 1
Convolutional Neural Networks: Part 1
 
Word representation: SVD, LSA, Word2Vec
Word representation: SVD, LSA, Word2VecWord representation: SVD, LSA, Word2Vec
Word representation: SVD, LSA, Word2Vec
 
Deep Learning For Speech Recognition
Deep Learning For Speech RecognitionDeep Learning For Speech Recognition
Deep Learning For Speech Recognition
 
Present simple tense
Present simple tensePresent simple tense
Present simple tense
 
Present simple tense
Present simple tensePresent simple tense
Present simple tense
 
Introduction and Starting ASP.NET MVC
Introduction and Starting ASP.NET MVCIntroduction and Starting ASP.NET MVC
Introduction and Starting ASP.NET MVC
 
Electronics projects
Electronics projectsElectronics projects
Electronics projects
 
ADO.NET Introduction
ADO.NET IntroductionADO.NET Introduction
ADO.NET Introduction
 
Azure machine learning tech mela
Azure machine learning tech melaAzure machine learning tech mela
Azure machine learning tech mela
 
Operations Management Study in Textured Jersy Lanka Limited
Operations Management Study in Textured Jersy Lanka LimitedOperations Management Study in Textured Jersy Lanka Limited
Operations Management Study in Textured Jersy Lanka Limited
 
Machine learning and azure ml studio
Machine learning and azure ml studioMachine learning and azure ml studio
Machine learning and azure ml studio
 
RC4&RC5
RC4&RC5RC4&RC5
RC4&RC5
 
L05 word representation
L05 word representationL05 word representation
L05 word representation
 
Machine learning and azure ml studio gabc
Machine learning and azure ml studio gabcMachine learning and azure ml studio gabc
Machine learning and azure ml studio gabc
 
Arduino Workshop Day 2
Arduino  Workshop Day 2Arduino  Workshop Day 2
Arduino Workshop Day 2
 

Semelhante a Deep Learning For Practitioners, lecture 2: Selecting the right applications deep learning

Unit one ppt of deeep learning which includes Ann cnn
Unit one ppt of  deeep learning which includes Ann cnnUnit one ppt of  deeep learning which includes Ann cnn
Unit one ppt of deeep learning which includes Ann cnnkartikaursang53
 
Deep learning short introduction
Deep learning short introductionDeep learning short introduction
Deep learning short introductionAdwait Bhave
 
Multi-modal sources for predictive modeling using deep learning
Multi-modal sources for predictive modeling using deep learningMulti-modal sources for predictive modeling using deep learning
Multi-modal sources for predictive modeling using deep learningSanghamitra Deb
 
"Solving Vision Tasks Using Deep Learning: An Introduction," a Presentation f...
"Solving Vision Tasks Using Deep Learning: An Introduction," a Presentation f..."Solving Vision Tasks Using Deep Learning: An Introduction," a Presentation f...
"Solving Vision Tasks Using Deep Learning: An Introduction," a Presentation f...Edge AI and Vision Alliance
 
ML crash course
ML crash courseML crash course
ML crash coursemikaelhuss
 
NLP and Deep Learning for non_experts
NLP and Deep Learning for non_expertsNLP and Deep Learning for non_experts
NLP and Deep Learning for non_expertsSanghamitra Deb
 
AI hype or reality
AI  hype or realityAI  hype or reality
AI hype or realityAwantik Das
 
Presentation 7.pptx
Presentation 7.pptxPresentation 7.pptx
Presentation 7.pptxShivam327815
 
Deep learning with tensorflow
Deep learning with tensorflowDeep learning with tensorflow
Deep learning with tensorflowCharmi Chokshi
 
Artificial Neural Networks for data mining
Artificial Neural Networks for data miningArtificial Neural Networks for data mining
Artificial Neural Networks for data miningALIZAIB KHAN
 
Big Data Analytics (ML, DL, AI) hands-on
Big Data Analytics (ML, DL, AI) hands-onBig Data Analytics (ML, DL, AI) hands-on
Big Data Analytics (ML, DL, AI) hands-onDony Riyanto
 
Handwritten Recognition using Deep Learning with R
Handwritten Recognition using Deep Learning with RHandwritten Recognition using Deep Learning with R
Handwritten Recognition using Deep Learning with RPoo Kuan Hoong
 
Artificial general intelligence (AGI) or Strong A.I.
Artificial general intelligence (AGI) or Strong A.I.Artificial general intelligence (AGI) or Strong A.I.
Artificial general intelligence (AGI) or Strong A.I.Pratap Dangeti
 

Semelhante a Deep Learning For Practitioners, lecture 2: Selecting the right applications deep learning (20)

Unit one ppt of deeep learning which includes Ann cnn
Unit one ppt of  deeep learning which includes Ann cnnUnit one ppt of  deeep learning which includes Ann cnn
Unit one ppt of deeep learning which includes Ann cnn
 
Deep learning short introduction
Deep learning short introductionDeep learning short introduction
Deep learning short introduction
 
Multi-modal sources for predictive modeling using deep learning
Multi-modal sources for predictive modeling using deep learningMulti-modal sources for predictive modeling using deep learning
Multi-modal sources for predictive modeling using deep learning
 
"Solving Vision Tasks Using Deep Learning: An Introduction," a Presentation f...
"Solving Vision Tasks Using Deep Learning: An Introduction," a Presentation f..."Solving Vision Tasks Using Deep Learning: An Introduction," a Presentation f...
"Solving Vision Tasks Using Deep Learning: An Introduction," a Presentation f...
 
ML crash course
ML crash courseML crash course
ML crash course
 
NLP and Deep Learning for non_experts
NLP and Deep Learning for non_expertsNLP and Deep Learning for non_experts
NLP and Deep Learning for non_experts
 
Deep learning for NLP
Deep learning for NLPDeep learning for NLP
Deep learning for NLP
 
AI hype or reality
AI  hype or realityAI  hype or reality
AI hype or reality
 
Presentation 7.pptx
Presentation 7.pptxPresentation 7.pptx
Presentation 7.pptx
 
tensorflow.pptx
tensorflow.pptxtensorflow.pptx
tensorflow.pptx
 
Deep learning with tensorflow
Deep learning with tensorflowDeep learning with tensorflow
Deep learning with tensorflow
 
presentation.ppt
presentation.pptpresentation.ppt
presentation.ppt
 
Lecture 1
Lecture 1Lecture 1
Lecture 1
 
lec1.ppt
lec1.pptlec1.ppt
lec1.ppt
 
Artificial Neural Networks for data mining
Artificial Neural Networks for data miningArtificial Neural Networks for data mining
Artificial Neural Networks for data mining
 
Artificial Neural Networks for Data Mining
Artificial Neural Networks for Data MiningArtificial Neural Networks for Data Mining
Artificial Neural Networks for Data Mining
 
Deep Learning
Deep LearningDeep Learning
Deep Learning
 
Big Data Analytics (ML, DL, AI) hands-on
Big Data Analytics (ML, DL, AI) hands-onBig Data Analytics (ML, DL, AI) hands-on
Big Data Analytics (ML, DL, AI) hands-on
 
Handwritten Recognition using Deep Learning with R
Handwritten Recognition using Deep Learning with RHandwritten Recognition using Deep Learning with R
Handwritten Recognition using Deep Learning with R
 
Artificial general intelligence (AGI) or Strong A.I.
Artificial general intelligence (AGI) or Strong A.I.Artificial general intelligence (AGI) or Strong A.I.
Artificial general intelligence (AGI) or Strong A.I.
 

Mais de ananth

Generative Adversarial Networks : Basic architecture and variants
Generative Adversarial Networks : Basic architecture and variantsGenerative Adversarial Networks : Basic architecture and variants
Generative Adversarial Networks : Basic architecture and variantsananth
 
Convolutional Neural Networks : Popular Architectures
Convolutional Neural Networks : Popular ArchitecturesConvolutional Neural Networks : Popular Architectures
Convolutional Neural Networks : Popular Architecturesananth
 
Overview of Convolutional Neural Networks
Overview of Convolutional Neural NetworksOverview of Convolutional Neural Networks
Overview of Convolutional Neural Networksananth
 
An Overview of Naïve Bayes Classifier
An Overview of Naïve Bayes Classifier An Overview of Naïve Bayes Classifier
An Overview of Naïve Bayes Classifier ananth
 
Mathematical Background for Artificial Intelligence
Mathematical Background for Artificial IntelligenceMathematical Background for Artificial Intelligence
Mathematical Background for Artificial Intelligenceananth
 
Search problems in Artificial Intelligence
Search problems in Artificial IntelligenceSearch problems in Artificial Intelligence
Search problems in Artificial Intelligenceananth
 
Introduction to Artificial Intelligence
Introduction to Artificial IntelligenceIntroduction to Artificial Intelligence
Introduction to Artificial Intelligenceananth
 
Machine Learning Lecture 3 Decision Trees
Machine Learning Lecture 3 Decision TreesMachine Learning Lecture 3 Decision Trees
Machine Learning Lecture 3 Decision Treesananth
 
Recurrent Neural Networks, LSTM and GRU
Recurrent Neural Networks, LSTM and GRURecurrent Neural Networks, LSTM and GRU
Recurrent Neural Networks, LSTM and GRUananth
 
An overview of Hidden Markov Models (HMM)
An overview of Hidden Markov Models (HMM)An overview of Hidden Markov Models (HMM)
An overview of Hidden Markov Models (HMM)ananth
 

Mais de ananth (10)

Generative Adversarial Networks : Basic architecture and variants
Generative Adversarial Networks : Basic architecture and variantsGenerative Adversarial Networks : Basic architecture and variants
Generative Adversarial Networks : Basic architecture and variants
 
Convolutional Neural Networks : Popular Architectures
Convolutional Neural Networks : Popular ArchitecturesConvolutional Neural Networks : Popular Architectures
Convolutional Neural Networks : Popular Architectures
 
Overview of Convolutional Neural Networks
Overview of Convolutional Neural NetworksOverview of Convolutional Neural Networks
Overview of Convolutional Neural Networks
 
An Overview of Naïve Bayes Classifier
An Overview of Naïve Bayes Classifier An Overview of Naïve Bayes Classifier
An Overview of Naïve Bayes Classifier
 
Mathematical Background for Artificial Intelligence
Mathematical Background for Artificial IntelligenceMathematical Background for Artificial Intelligence
Mathematical Background for Artificial Intelligence
 
Search problems in Artificial Intelligence
Search problems in Artificial IntelligenceSearch problems in Artificial Intelligence
Search problems in Artificial Intelligence
 
Introduction to Artificial Intelligence
Introduction to Artificial IntelligenceIntroduction to Artificial Intelligence
Introduction to Artificial Intelligence
 
Machine Learning Lecture 3 Decision Trees
Machine Learning Lecture 3 Decision TreesMachine Learning Lecture 3 Decision Trees
Machine Learning Lecture 3 Decision Trees
 
Recurrent Neural Networks, LSTM and GRU
Recurrent Neural Networks, LSTM and GRURecurrent Neural Networks, LSTM and GRU
Recurrent Neural Networks, LSTM and GRU
 
An overview of Hidden Markov Models (HMM)
An overview of Hidden Markov Models (HMM)An overview of Hidden Markov Models (HMM)
An overview of Hidden Markov Models (HMM)
 

Último

RadioAdProWritingCinderellabyButleri.pdf
RadioAdProWritingCinderellabyButleri.pdfRadioAdProWritingCinderellabyButleri.pdf
RadioAdProWritingCinderellabyButleri.pdfgstagge
 
Heart Disease Classification Report: A Data Analysis Project
Heart Disease Classification Report: A Data Analysis ProjectHeart Disease Classification Report: A Data Analysis Project
Heart Disease Classification Report: A Data Analysis ProjectBoston Institute of Analytics
 
科罗拉多大学波尔得分校毕业证学位证成绩单-可办理
科罗拉多大学波尔得分校毕业证学位证成绩单-可办理科罗拉多大学波尔得分校毕业证学位证成绩单-可办理
科罗拉多大学波尔得分校毕业证学位证成绩单-可办理e4aez8ss
 
Effects of Smartphone Addiction on the Academic Performances of Grades 9 to 1...
Effects of Smartphone Addiction on the Academic Performances of Grades 9 to 1...Effects of Smartphone Addiction on the Academic Performances of Grades 9 to 1...
Effects of Smartphone Addiction on the Academic Performances of Grades 9 to 1...limedy534
 
GA4 Without Cookies [Measure Camp AMS]
GA4 Without Cookies [Measure Camp AMS]GA4 Without Cookies [Measure Camp AMS]
GA4 Without Cookies [Measure Camp AMS]📊 Markus Baersch
 
办理学位证纽约大学毕业证(NYU毕业证书)原版一比一
办理学位证纽约大学毕业证(NYU毕业证书)原版一比一办理学位证纽约大学毕业证(NYU毕业证书)原版一比一
办理学位证纽约大学毕业证(NYU毕业证书)原版一比一fhwihughh
 
Predictive Analysis for Loan Default Presentation : Data Analysis Project PPT
Predictive Analysis for Loan Default  Presentation : Data Analysis Project PPTPredictive Analysis for Loan Default  Presentation : Data Analysis Project PPT
Predictive Analysis for Loan Default Presentation : Data Analysis Project PPTBoston Institute of Analytics
 
NLP Data Science Project Presentation:Predicting Heart Disease with NLP Data ...
NLP Data Science Project Presentation:Predicting Heart Disease with NLP Data ...NLP Data Science Project Presentation:Predicting Heart Disease with NLP Data ...
NLP Data Science Project Presentation:Predicting Heart Disease with NLP Data ...Boston Institute of Analytics
 
PKS-TGC-1084-630 - Stage 1 Proposal.pptx
PKS-TGC-1084-630 - Stage 1 Proposal.pptxPKS-TGC-1084-630 - Stage 1 Proposal.pptx
PKS-TGC-1084-630 - Stage 1 Proposal.pptxPramod Kumar Srivastava
 
Call Us ➥97111√47426🤳Call Girls in Aerocity (Delhi NCR)
Call Us ➥97111√47426🤳Call Girls in Aerocity (Delhi NCR)Call Us ➥97111√47426🤳Call Girls in Aerocity (Delhi NCR)
Call Us ➥97111√47426🤳Call Girls in Aerocity (Delhi NCR)jennyeacort
 
RS 9000 Call In girls Dwarka Mor (DELHI)⇛9711147426🔝Delhi
RS 9000 Call In girls Dwarka Mor (DELHI)⇛9711147426🔝DelhiRS 9000 Call In girls Dwarka Mor (DELHI)⇛9711147426🔝Delhi
RS 9000 Call In girls Dwarka Mor (DELHI)⇛9711147426🔝Delhijennyeacort
 
Semantic Shed - Squashing and Squeezing.pptx
Semantic Shed - Squashing and Squeezing.pptxSemantic Shed - Squashing and Squeezing.pptx
Semantic Shed - Squashing and Squeezing.pptxMike Bennett
 
NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...
NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...
NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...Amil Baba Dawood bangali
 
Advanced Machine Learning for Business Professionals
Advanced Machine Learning for Business ProfessionalsAdvanced Machine Learning for Business Professionals
Advanced Machine Learning for Business ProfessionalsVICTOR MAESTRE RAMIREZ
 
NLP Project PPT: Flipkart Product Reviews through NLP Data Science.pptx
NLP Project PPT: Flipkart Product Reviews through NLP Data Science.pptxNLP Project PPT: Flipkart Product Reviews through NLP Data Science.pptx
NLP Project PPT: Flipkart Product Reviews through NLP Data Science.pptxBoston Institute of Analytics
 
Data Factory in Microsoft Fabric (MsBIP #82)
Data Factory in Microsoft Fabric (MsBIP #82)Data Factory in Microsoft Fabric (MsBIP #82)
Data Factory in Microsoft Fabric (MsBIP #82)Cathrine Wilhelmsen
 
Indian Call Girls in Abu Dhabi O5286O24O8 Call Girls in Abu Dhabi By Independ...
Indian Call Girls in Abu Dhabi O5286O24O8 Call Girls in Abu Dhabi By Independ...Indian Call Girls in Abu Dhabi O5286O24O8 Call Girls in Abu Dhabi By Independ...
Indian Call Girls in Abu Dhabi O5286O24O8 Call Girls in Abu Dhabi By Independ...dajasot375
 
Easter Eggs From Star Wars and in cars 1 and 2
Easter Eggs From Star Wars and in cars 1 and 2Easter Eggs From Star Wars and in cars 1 and 2
Easter Eggs From Star Wars and in cars 1 and 217djon017
 
原版1:1定制南十字星大学毕业证(SCU毕业证)#文凭成绩单#真实留信学历认证永久存档
原版1:1定制南十字星大学毕业证(SCU毕业证)#文凭成绩单#真实留信学历认证永久存档原版1:1定制南十字星大学毕业证(SCU毕业证)#文凭成绩单#真实留信学历认证永久存档
原版1:1定制南十字星大学毕业证(SCU毕业证)#文凭成绩单#真实留信学历认证永久存档208367051
 

Último (20)

RadioAdProWritingCinderellabyButleri.pdf
RadioAdProWritingCinderellabyButleri.pdfRadioAdProWritingCinderellabyButleri.pdf
RadioAdProWritingCinderellabyButleri.pdf
 
Heart Disease Classification Report: A Data Analysis Project
Heart Disease Classification Report: A Data Analysis ProjectHeart Disease Classification Report: A Data Analysis Project
Heart Disease Classification Report: A Data Analysis Project
 
科罗拉多大学波尔得分校毕业证学位证成绩单-可办理
科罗拉多大学波尔得分校毕业证学位证成绩单-可办理科罗拉多大学波尔得分校毕业证学位证成绩单-可办理
科罗拉多大学波尔得分校毕业证学位证成绩单-可办理
 
Effects of Smartphone Addiction on the Academic Performances of Grades 9 to 1...
Effects of Smartphone Addiction on the Academic Performances of Grades 9 to 1...Effects of Smartphone Addiction on the Academic Performances of Grades 9 to 1...
Effects of Smartphone Addiction on the Academic Performances of Grades 9 to 1...
 
GA4 Without Cookies [Measure Camp AMS]
GA4 Without Cookies [Measure Camp AMS]GA4 Without Cookies [Measure Camp AMS]
GA4 Without Cookies [Measure Camp AMS]
 
办理学位证纽约大学毕业证(NYU毕业证书)原版一比一
办理学位证纽约大学毕业证(NYU毕业证书)原版一比一办理学位证纽约大学毕业证(NYU毕业证书)原版一比一
办理学位证纽约大学毕业证(NYU毕业证书)原版一比一
 
Predictive Analysis for Loan Default Presentation : Data Analysis Project PPT
Predictive Analysis for Loan Default  Presentation : Data Analysis Project PPTPredictive Analysis for Loan Default  Presentation : Data Analysis Project PPT
Predictive Analysis for Loan Default Presentation : Data Analysis Project PPT
 
NLP Data Science Project Presentation:Predicting Heart Disease with NLP Data ...
NLP Data Science Project Presentation:Predicting Heart Disease with NLP Data ...NLP Data Science Project Presentation:Predicting Heart Disease with NLP Data ...
NLP Data Science Project Presentation:Predicting Heart Disease with NLP Data ...
 
PKS-TGC-1084-630 - Stage 1 Proposal.pptx
PKS-TGC-1084-630 - Stage 1 Proposal.pptxPKS-TGC-1084-630 - Stage 1 Proposal.pptx
PKS-TGC-1084-630 - Stage 1 Proposal.pptx
 
Call Us ➥97111√47426🤳Call Girls in Aerocity (Delhi NCR)
Call Us ➥97111√47426🤳Call Girls in Aerocity (Delhi NCR)Call Us ➥97111√47426🤳Call Girls in Aerocity (Delhi NCR)
Call Us ➥97111√47426🤳Call Girls in Aerocity (Delhi NCR)
 
Deep Generative Learning for All - The Gen AI Hype (Spring 2024)
Deep Generative Learning for All - The Gen AI Hype (Spring 2024)Deep Generative Learning for All - The Gen AI Hype (Spring 2024)
Deep Generative Learning for All - The Gen AI Hype (Spring 2024)
 
RS 9000 Call In girls Dwarka Mor (DELHI)⇛9711147426🔝Delhi
RS 9000 Call In girls Dwarka Mor (DELHI)⇛9711147426🔝DelhiRS 9000 Call In girls Dwarka Mor (DELHI)⇛9711147426🔝Delhi
RS 9000 Call In girls Dwarka Mor (DELHI)⇛9711147426🔝Delhi
 
Semantic Shed - Squashing and Squeezing.pptx
Semantic Shed - Squashing and Squeezing.pptxSemantic Shed - Squashing and Squeezing.pptx
Semantic Shed - Squashing and Squeezing.pptx
 
NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...
NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...
NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...
 
Advanced Machine Learning for Business Professionals
Advanced Machine Learning for Business ProfessionalsAdvanced Machine Learning for Business Professionals
Advanced Machine Learning for Business Professionals
 
NLP Project PPT: Flipkart Product Reviews through NLP Data Science.pptx
NLP Project PPT: Flipkart Product Reviews through NLP Data Science.pptxNLP Project PPT: Flipkart Product Reviews through NLP Data Science.pptx
NLP Project PPT: Flipkart Product Reviews through NLP Data Science.pptx
 
Data Factory in Microsoft Fabric (MsBIP #82)
Data Factory in Microsoft Fabric (MsBIP #82)Data Factory in Microsoft Fabric (MsBIP #82)
Data Factory in Microsoft Fabric (MsBIP #82)
 
Indian Call Girls in Abu Dhabi O5286O24O8 Call Girls in Abu Dhabi By Independ...
Indian Call Girls in Abu Dhabi O5286O24O8 Call Girls in Abu Dhabi By Independ...Indian Call Girls in Abu Dhabi O5286O24O8 Call Girls in Abu Dhabi By Independ...
Indian Call Girls in Abu Dhabi O5286O24O8 Call Girls in Abu Dhabi By Independ...
 
Easter Eggs From Star Wars and in cars 1 and 2
Easter Eggs From Star Wars and in cars 1 and 2Easter Eggs From Star Wars and in cars 1 and 2
Easter Eggs From Star Wars and in cars 1 and 2
 
原版1:1定制南十字星大学毕业证(SCU毕业证)#文凭成绩单#真实留信学历认证永久存档
原版1:1定制南十字星大学毕业证(SCU毕业证)#文凭成绩单#真实留信学历认证永久存档原版1:1定制南十字星大学毕业证(SCU毕业证)#文凭成绩单#真实留信学历认证永久存档
原版1:1定制南十字星大学毕业证(SCU毕业证)#文凭成绩单#真实留信学历认证永久存档
 

Deep Learning For Practitioners, lecture 2: Selecting the right applications deep learning

  • 1. Deep  Learning  For  Prac//oners     Lecture  2:  Which  applica/ons  benefit  from   deep  learning?   Anantharaman  Narayana  Iyer     deeplearning.ananth@gmail.com   17th  June  2014   Note:  Notes  that  contain  code  examples  for  these  slides  and   detailed  analysis  will  be  published  separately  later.  
  • 2. Review  of  previous  lecture   •  Deep  learning  as  a  major  machine  learning  discipline   has  received  phenomenal  aNen/on  of  late  due  to:   –  Breakthrough  results  reported  by  the  research   community    for  certain  class  of  applica/ons,  beNering   the  current  state  of  the  art   –  Substan/al  investments  by  technology  companies   such  as:  Google,  Facebook,  MicrosoU,  IBM     •  While  there  is  no  single  unique  architecture,  deep   networks  are  typically  built  using  some  variant  of   Autoencoders  or  Restricted  Boltzmann  Machines   with  key  characteris/cs  of:   –  Deep  architecture:  Mul/ple  layers  performing   complex,  nonlinear  computa/ons,  cascading  the   layerwise  outputs.   –  Automated  feature  extrac/on:  each  layer  produces  as   its  output  an  abstracted  form  of  its  inputs  (e.g.  Edges   from  raw  pixels).  One  may  add  a  classifier  layer  (e.g   SVM)  on  top  of  the  abstracted  features  and  can  view   the  classifica/on  as  being  done  on  the  most  abstract   features  automa/cally  generated  by  the  system.  (An   example  with  code  illustrated  in  the  next  lecture)  
  • 3. Looking  through  the  prac//oner’s  prism   •  To  address  real  world  problems,   prac//oners  need  to  be  aware  of  where   deep  learning  yields  best  results,  prac/cal   considera/ons,  limita/ons  and  when  not  to   use  it.     •  This  requires  looking  at  the  research  results   and  other  claims  from  a  prac/cal   perspec/ve  and  stay  clear  of  common   misconcep/ons.  
  • 4. “If  all  you  have  is  a  hammer  everything  looks  as  a  nail”   •  Deep  learning  has  proved  its  poten/al  in  some  applica/on  domains  (e.g.   Computer  Vision,  Speech  recogni/on),  holds  early  promise  in  several  other  areas   (e.g  Natural  Language  Processing)  but  this  is  not  a  universal  tool  to  provide  the   best  result  for  “any”  AI  task.   •  When  does  it  have  the  poten/al  to  perform  best?   –  When  structure  of  the  problem  being  solved  naturally  maps  to  a  mul/  layer   architecture   •  If  the  problem  we  are  trying  to  solve  can  be  decomposed  in  to  processing  hierarchical   abstract  features  and  these  features  are  derivable  from  the  input  data  through  a  set  of   poten/ally  nonlinear  transforma/ons,  deep  learning  based  solu/on  might  be  effec/ve.     •  As  a  corollary,  problems  that  don’t  exhibit  a  mul/  layer  structure  may  not  see  much   incremental  benefit  compared  to  tradi/onal  methods   –  Data  availability   •  While  tradi/onal  architectures  require  expert  designed  features,  deep  learning  systems   automa/cally  learn  these  features,  given  the  raw  input.   •  In  order  to  learn  the  features,  extensive,  unsupervised  pretraining  using  large  volumes  of   data  is  oUen  required.  Hence  any  advanced  solu/on  based  on  deep  learning  is  likely  to   require  availability  of  such  data.  
  • 5. “More  data  or  beNer  models?”   •  Data  Vs  Algorithm:  research  shows  that   training  a  system  with  more  data,  the   performance  asympto/cally  approaches   same  levels  regardless  of  the  model.   •  One  may  be  led  to  believe  that  shallow   networks,  trained  with  huge  data  might   equal  the  performance  of  deep  networks.     –  Unfortunately,  much  of  the  available  data  in  the   web  is  unlabeled  and  without  an  effec/ve   unsupervised  training  model,  the  data  is  not   useful.  Deep  networks  with  unsupervised   pretraining  phase,  can  leverage  the  data  beNer.   •  Another  no/on  could  be  that  any   algorithm  or  model  selec/on  for  a  deep   network  is  good  enough  if  you  give  it  a   huge  volume  of  data.   –  Choosing  an  op/mal  algorithm  and  design  is   very  cri/cal  as  deep  networks  are  resource   heavy  due  to  mul/ple  layers  and  weights.  A   good  intui/on  on  the  problem  structure  is   important  to  make  right  choices  of  the  model.  
  • 6. Automated  Feature  Learning  and  data  preprocessing   Though  deep  learning  systems  extract  features  automa/cally,  the  task  of  data   preprocessing  is  s/ll  non-­‐trivial.   –  The  input  data  should  be  complete  enough  so  that  the  features  relevant  for  the  given   problem  can  be  extracted.   •  Consider  the  example  of  detec/ng  anomalies  in  the  opera/on  of  a  nuclear  reactor.  The   input  to  be  given  to  a  deep  learning  system  should  include  signals  from  all  the  relevant   sensors  and  missing  any  of  them  may  result  in  inadequate  performance   –  The  op/mum  size  of  the  input  data  adequate  for  the  job  needs  to  be  determined.   •  Suppose  we  need  to  perform  face  detec/on,  given  the  input  images.  What  should  be  the   right  input  size?  Should  it  be  10  x  10  or  100  x  100  pixels?  High  dimensionality  increases  the   model  parameters  substan/ally,  requiring  more  compute  resources.   –  Input  vector  representa/on  must  be  determined   •  E.g,  for  an  NLP  problem,  words  from  a  vocabulary  V  may  be  represented  in  “one-­‐hot”  form   where  each  word  in  V  is  represented  by  a  posi/on.  Here,  the  number  of  features  for  a  given   word  w  equals  the  size  of    the  vocabulary  |V|  and  a  sentence  with  k  words  will  be   represented  as  k  *  |V|  sized  input  vector.  When  the  size  of  vocabulary  becomes  large  (say   over  10000  words),  this  representa/on  increases  the  dimensionality  substan/ally.   –  For  many  problems,  data  cleaning  and  preprocessing  are  s/ll  required   •  E.g.  For  many  NLP  problems,  beNer  performance  may  be  obtained  easier  through  some   preprocessing  steps  (such  as  stopword  removal,  stemming  etc)  rather  than  lehng  the  deep   learning  system  handle  the  data  in  its  raw  form.