why does averaging work n.
Download
Skip this Video
Loading SlideShow in 5 Seconds..
Why does averaging work? PowerPoint Presentation
Download Presentation
Why does averaging work?

Loading in 2 Seconds...

play fullscreen
1 / 45

Why does averaging work? - PowerPoint PPT Presentation


  • 67 Views
  • Uploaded on

Why does averaging work?. Yoav Freund AT&T Labs. Plan of talk. Generative vs. non-generative modeling Boosting Boosting and over-fitting Bagging and over-fitting Applications. Male. Human Voice. Female. Toy Example. Computer receives telephone call Measures Pitch of voice

loader
I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.
capcha
Download Presentation

PowerPoint Slideshow about 'Why does averaging work?' - jonah


An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript
why does averaging work

Why does averaging work?

Yoav Freund AT&T Labs

plan of talk
Plan of talk
  • Generative vs. non-generative modeling
  • Boosting
  • Boosting and over-fitting
  • Bagging and over-fitting
  • Applications
toy example

Male

Human

Voice

Female

Toy Example
  • Computer receives telephone call
  • Measures Pitch of voice
  • Decides gender of caller
generative modeling

mean1

mean2

var1

var2

Probability

Generative modeling

Voice Pitch

ill behaved data

mean2

mean1

Probability

No. of mistakes

Ill-behaved data

Voice Pitch

traditional statistics vs machine learning

Machine Learning

Decision

Theory

Statistics

Traditional Statistics vs. Machine Learning

Predictions

Actions

Data

Estimated

world state

another example
Another example
  • Two dimensional data (example: pitch, volume)
  • Generative model: logistic regression
  • Discriminative model: separating line ( “perceptron” )
slide9

Model:

Find W to maximize

w

slide10

w

Model:

Find W to maximize

adaboost as gradient descent
Adaboost as gradient descent
  • Discriminator class: a linear discriminator in the space of “weak hypotheses”
  • Original goal: find hyper plane with smallest number of mistakes
    • Known to be an NP-hard problem (no algorithm that runs in time polynomial in d, where d is the dimension of the space)
  • Computational method: Use exponential loss as a surrogate, perform gradient descent.
  • Unforeseen benefit: Out-of-sample error improves even after in-sample error reaches zero.
margins view

Prediction =

Margin =

-

Cumulative # examples

w

-

+

+

+

-

Mistakes

Correct

-

+

+

+

-

-

-

-

+

Correct

Mistakes

Margin

Margins view

Project

adaboost et al

Adaboost =

Logitboost

Brownboost

0-1 loss

Margin

Mistakes

Correct

Adaboost et al.

Loss

one coordinate at a time
One coordinate at a time
  • Adaboost performs gradient descent on exponential loss
  • Adds one coordinate (“weak learner”) at each iteration.
  • Weak learning in binary classification = slightly better than random guessing.
  • Weak learning in regression – unclear.
  • Uses example-weights to communicate the gradient direction to the weak learner
  • Solves a computational problem
curious phenomenon
Curious phenomenon

Boosting decision trees

Using <10,000 training examples we fit >2,000,000 parameters

theorem

Fraction of training example

with small margin

Probability of mistake

Size of training sample

VC dimension of weak rules

Theorem

Schapire, Freund, Bartlett & Lee

Annals of stat. 98

For any convex combination and any threshold

No dependence on number

of weak rules

that are combined!!!

a metric space of classifiers

d

f

g

d(f,g) = P( f(x) = g(x) )

A metric space of classifiers

Classifier space

Example Space

Neighboring models make similar predictions

confidence zones

Certainly

Certainly

Bagged variants

No. of mistakes

Majority vote

Confidence zones

Voice Pitch

bagging
Bagging
  • Averaging over all good models increases stability (decreases dependence on particular sample)
  • If clear majority the prediction is very reliable(unlike Florida)
  • Bagging is a randomized algorithm for sampling the best model’s neighborhood
  • Ignoring computation, averaging of best models can be done deterministically, and yield provable bounds.
a pseudo bayesian method

set

;

Weights:

Log odds:

+1

l(x)>+

Prediction(x) =

0

l(x)< -

-1

Freund, Mansour, Schapire 2000

A pseudo-Bayesian method

Define a prior distribution over rulesp(h)

Get m training examples,

For each h : err(h) = (no. of mistakes h makes)/m

else

Insufficient data

applications of boosting
Applications of Boosting
  • Academic research
  • Applied research
  • Commercial deployment
academic research
Academic research

% test error rates

applied research

Schapire, Singer, Gorin 98

Applied research
  • “AT&T, How may I help you?”
  • Classify voice requests
  • Voice -> text -> category
  • Fourteen categories

Area code, AT&T service, billing credit, calling card, collect, competitor, dial assistance, directory, how to dial, person to person, rate, third party, time charge ,time

examples
Examples
  • Yes I’d like to place a collect call long distance please
  • Operator I need to make a call but I need to bill it to my office
  • Yes I’d like to place a call on my master card please
  • I just called a number in Sioux city and I musta rang the wrong number because I got the wrong party and I would like to have that taken off my bill
  • collect
  • third party
  • calling card
  • billing credit
weak rules generated by boostexter

Word occurs

Word does

not occur

Weak rules generated by “boostexter”

Third party

Collect call

Calling card

Category

Weak Rule

results
Results
  • 7844 training examples
    • hand transcribed
  • 1000 test examples
    • hand / machine transcribed
  • Accuracy with 20% rejected
    • Machine transcribed: 75%
    • Hand transcribed: 90%
commercial deployment

Freund, Mason, Rogers, Pregibon, Cortes 2000

Commercial deployment
  • Distinguish business/residence customers
  • Using statistics from call-detail records
  • Alternating decision trees
    • Similar to boosting decision trees, more flexible
    • Combines very simple rules
    • Can over-fit, cross validation used to stop
massive datasets
Massive datasets
  • 260M calls / day
  • 230M telephone numbers
  • Label unknown for ~30%
  • Hancock: software for computing statistical signatures.
  • 100K randomly selected training examples,
  • ~10K is enough
  • Training takes about 2 hours.
  • Generated classifier has to be both accurateand efficient
business impact
Business impact
  • Increased coverage from 44% to 56%
  • Accuracy ~94%
  • Saved AT&T 15M$ in the year 2000 in operations costs and missed opportunities.
summary
Summary
  • Boosting is a computational method for learning accurate classifiers
  • Resistance to over-fit explained by margins
  • Underlying explanation – large “neighborhoods” of good classifiers
  • Boosting has been applied successfully to a variety of classification problems
please come talk with me
Please come talk with me
  • Questions welcome!
  • Data, even better!!
  • Potential collaborations, yes!!!
  • Yoav@research.att.com
  • www.predictiontools.com