Machine Learning What, how, why? - GitHub Pages

Machine LearningWhat, how, why?

Rémi Emonet (@remiemonet)2015-09-30

Web En Vert

$ whoami

IBM BlueMix (Watson)

$ whoami

Software EngineerResearcher: machine learning, computer visionTeacher: web technologies, computing literacyGeek: deck.js slides, isochrones, …

You are shrewd, skeptical and restrained.You are independent: you have a strong desire to have time to yourself. You arecalm-seeking: you prefer activities that are quiet, calm, and safe. And you arephilosophical: you are open to and intrigued by new ideas and love to explorethem.Experiences that give a sense of prestige hold some appeal to you.You are relatively unconcerned with both tradition and taking pleasure in life.You care more about making your own path than following what others havedone. And you prefer activities with a purpose greater than just personalenjoyment.

4 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

What is Machine Learning?

5 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Machine Learning Basic PrincipleGiven a dataset {y , x , … , x }

Optimize the likelihood functionL = n(w, t , d)log p(w, t ∣z)p(z, t ∣d)

Or using a sparse regularizationL − λ KL(U ∣∣p(ts∣z, d))

By using a Gibbs Samplerp(W , at ∣o = o, O ) =

i i1 ip i=1n

d

∑w

∑ta

∑ a

z

∑ts

∑ r s

sparse

d

∑z

∑

ji ji ji−ji

N (w , rt , z ) + η(w , rt )∑w ,rt′ ′ ( obs−ji ′ ′

ji′ ′ )

N (W , rt , z ) + η(W , rt )obs−ji

ji ji ji ji ji

6 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Machine Learningin the Wild

Which One of These ServicesUses Machine Learning?

7 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

8 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Machine Learning in Future Tech?

9 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

What is Machine Learning?an example motivation

Challenge: Which Iris Species?

11 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

⇒Sepal length: 5.1Sepal width: 2.5Petal length: 4.2Petal width: 1.0

Expected Label: “Iris Setosa”

Feature Extraction

Analysis and Program Writing

IFTTT...

14 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Predictive Machine LearningInstead of writing a program that solves a task,We

collect labeled data: input/output pairs1.

automatically generate a programthat computes an output for each new input

2.

profit!3.

The machine learns to generalizefrom a limited number of examples,

like humans do.

15 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Different Types of TasksSupervised learning: some labels are known

classification: find the label of an exampleregression: find the target value

Unsupervised learning: no labelsclustering: group things togetherpattern mining: find recurrent eventsanomaly detection: find “outliers”

16 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

The Principle of “Overfitting”

17 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

A Lot of Different MethodsAlso called “models”

linear regression, logistic regression, SVM, kernel SVM, neural networks, k-meansclustering, collaborative filtering, bayesian networks, expectation maximization,belief propagation, multiple kernel learning, metric learning, transfer learning,decision trees, gaussian processes, random forests, boosting, ...

For different contextsdifferent tasksdifferent nature of datadifferent suppositions on the datadifferent amount of data

18 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Different Ways to StartUse a product that uses ML

e.g. adwords, ibm bluemix, …

Use a prediction APIsend your data to the serviceget API to process new inputse.g., google pred. API, prediction.io, ...

Dive into machine learning…

19 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Into Machine LearningUsing libraries

libraries exist in most languagesmost models already implementedtest different methods with different parameters

Learning machine learningmany online coursesget deeper understanding

20 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Does Machine Learning ActuallyMatter?

21 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Example: The Netflix Challenge

FAIR

22 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

23 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Example: Facebook AI ResearchDirector: Yann LeCunScientific Leads

Léon BouttouRob FergusFlorent Perronnin

fr rest

24 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Data, Data, Data

25 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Data isMachine Learning's Fuel

data === power

Getting Data?

28 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Getting Data?Collect from your services/applicationsDo it yourselfPay some people you knowUse crowd-sourcing,e.g., Amazon Mechanical Turk (MTurk)Find existing datasets (open data, etc)Work for/with a “data rich” companyCreate your “intermediation” business

29 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

What Can It DoFor Me

30 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

SearchGoogle Search, Bing, etc

31 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

AdvertisingAdWords, etc

32 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

RecommendationsNetflix, Amazon, Youtube,

app Stores, etc

33 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Text Translation

34 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Optical Character Recognition(postcodes, checks, book scans,

etc)

35 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Visual Recognition (objects,plants, animals, etc)

36 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Face DetectionSmile Detection

(embedded in Cameras)

37 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Face Identification(Picasa, Facebook, etc)

38 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Kinect Controller

39 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Self Driving Cars

40 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Voice Recognitionand Synthesis

(GoogleNow, Siri, Cortana)

41 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Sound Recognition(birds, underwater sounds,

safety, etc)

42 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Fraud Detection(Banking, Websites, etc)

43 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Automated Trading…

44 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Customer/Person ProfilingBlueMix Watson, etc

45 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Adaptive Websites(automated A/B testing)

46 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

The “Big Data” Hype

and much more...

48 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

Where Will it Stop?

Singularity?

Thanks! Questions?

twitter: @remiemonetweb/email: http://home.heeere.comRecommended Links:

comprehensive introduction to ML modelsscikit learn (python)... 50 / 50 − Rémi Emonet (@remiemonet) − Machine Learning What, how, why?

tatadbb

-Porsupah-

ali eminov

efilpera

jannekestaaks

Asa-moya

francoisjouffroy

JD Hancock