lecture 01: machine learning for language technology - introduction

Lecture 1: IntroductionLast updated: 2 Sept 2013

Uppsala University

Department of Linguistics and Philology, September 2013

1 Lecture 1: Introduction

Machine Learning for Language Technology(Schedule)

http://stp.lingfil.uu.se/~santinim/ml/MachineLearning_fall2013.htm

Lecture 1: Introduction2

Practical Information

Reading list, assignments, exams, etc.

Most slides from previous courses (i.e. from E. Alpaydin and J. Nivre) with adaptations


Course web pages: http://stp.lingfil.uu.se/~santinim/ml/ml_fall2013.pdf


Contact details: [email protected] ([email protected])

http://stp.lingfil.uu.se/~santinim/ml/ml_fall2013.pdf



mailto:[email protected]


About the Course Introduction to machine learning Focus on methods used in Language Technology and

NLP Decision trees and nearest neighbor methods (Lecture 2) Linear models –The Weka ML package (Lectures 3 and 4) Ensemble methods – Structured Predictions (Lectures 5 and

6) Text Mining and Big Data - R and Rapid Miner(Lecture 7) Unsupervised learning (clustering) (Lecture 8, Magnus

Rosell)

Builds on Statistical Methods in NLP Mostly discriminative methods Generative probability models covered in first course


Digression: Generative vs. Discriminative Methods

A generative method only applies to probabilistic models. A model is generative if it gives us the model of the joint distribution of x and y together (P (Y , X ). It is called generative because you can generate with the correct probability distribution data points.

Conditional methods model the conditional distribution of the output given the input: P (Y | X ).

Discriminative methods do not model probabilities at all, but they map the input to the output directly.


Compulsory Reading List Main textbooks:

1. Ethem Alpaydin. 2010. Introduction to Machine Learning. Second Edition. MIT Press (free online version)

2. Hal Daumé III. 2012. A Course in Machine Learning (free online version)

3. Ian H. Witten, Eibe Frank Data Mining: Practical Machine Learning Tools and Techniques, Second Edition (free online version)

Additional material :1. Dietterich, T. G. (2000). Ensemble Methods in Machine

Learning. In J. Kittler and F. Roli (Ed.) First International Workshop on Multiple Classifier Systems

2. Michael Collins. 2002. Discriminative Training Methods for Hidden Markov Models: Theory and Experiments with Perceptron Algorithms.In Proceedings of the 2002 Conference on Empirical Methods in Natural Language Processing

3. Hanna M. Wallach. 2004. Conditional Random Fields: An Introduction.Technical Report MS-CIS-04-21. Department of Computer and Information Science, University of Pennsylvania.

http://www.realtechsupport.org/UB/MRIII/papers/MachineLearning/Alppaydin_MachineLearning_2010.pdf

http://ciml.info/dl/v0_8/ciml-v0_8-all.pdf

http://home.etf.rs/~vm/os/dmsw/Morgan.Kaufman.Publishers.Weka.2nd.Edition.2005.Elsevier.pdf


Optional Reading Hal Daumé III, John Langford, Daniel Marcu

(2005) Search-Based Structured Prediction as Classification. NIPS Workshop on Advances in Structured Learning for Text and Speech Processing (ASLTSP).


Assignments and Examination

Three Assignments: Decision trees and nearest neighbors Perceptron learning Clustering

General Info: No lab sessions, supervision by email Reports and Assignments 1 & 2 must be submitted to

[email protected] Report and Assignment 3 must be submitted to [email protected]

Examination: Written report submitted for each assignment All three assignments necessary to pass the course Grade determined by majority grade on assignments

mailto:[email protected]


Practical Organization Marina Santini (1-7); Magnus Rosell (8) 45min + 15 min break

Lectures on Course webpage and SlideShare Email all your questions to me:

[email protected] Video Recordings of the previous ML course:

http://stp.lingfil.uu.se/~nivre/master/ml.html

Send me an email, so I make sure that I have all the correct email addresses to [email protected]

http://stp.lingfil.uu.se/~nivre/master/ml.html


Schedule: http://stp.lingfil.uu.se/~santinim/ml/ml_fall2013.pdf




What is Machine Learning?

Introduction to:•Classification•Regression•Supervised Learning•Unsupervised Learning•Reinforcement Learning


What is Machine Learning Machine learning is programming computers

to optimize a performance criterion for some task using example data or past experience

Why learning? No known exact method – vision, speech

recognition, robotics, spam filters, etc. Exact method too expensive – statistical physics Task evolves over time – network routing

Compare: No need to use machine learning for computing

payroll… we just need an algorithm


Machine Learning – Data Mining – Artificial Intelligence – Statistics Machine Learning: creation of a model that uses training

data or past experience

Data Mining: application of learning methods to large datasets (ex. physics, astronomy, biology, etc.) Text mining = machine learning applied to unstructured textual

data (ex. sentiment analyisis, social media monitoring, etc. Text Mining, Wikipedia)

Artificial intelligence: a model that can adapt to a changing environment.

Statistics: Machine learning uses the theory of statistics in building mathematical models, because the core task is making inference from a sample.

http://en.wikipedia.org/wiki/Text_mining

http://en.wikipedia.org/wiki/Text_mining


The bio-cognitive analogy Imagine that a learning algorithm as a single

neuron. This neuron receives input from other

neurons, one for each input feature. The strength of these inputs are the feature

values. Each input has a weight and the neuron

simply sums up all the weighted inputs. Based on this sum, the neuron decides

whether to “fire” or not. Firing is interpreted as being a positive example and not firing is interpreted as being a negative example.


Elements of Machine Learning1. Generalization:

Generalize from specific examples Based on statistical inference

2. Data: Training data: specific examples to learn from Test data: (new) specific examples to assess

performance

3. Models: Theoretical assumptions about the task/domain Parameters that can be inferred from data

4. Algorithms: Learning algorithm: infer model (parameters) from data Inference algorithm: infer predictions from model


Types of Machine Learning Association Supervised Learning

Classification Regression

Unsupervised Learning Reinforcement Learning


Learning Associations Basket analysis:

P (Y | X ) probability that somebody who buys X also buys Y where X and Y are products/services

Example: P ( chips | beer ) = 0.7


Classification

Example: Credit scoring

Differentiating between low-risk and high-risk customers from their income and savings

Discriminant: IF income > θ1 AND savings > θ2 THEN low-risk ELSE high-risk


Classification in NLP

Binary classification: Spam filtering (spam vs. non-spam) Spelling error detection (error vs. non error)

Multiclass classification: Text categorization (news, economy, culture,

sport, ...) Named entity classification (person, location,

organization, ...) Structured prediction:

Part-of-speech tagging (classes = tag sequences) Syntactic parsing (classes = parse trees)

Regression

Example: Price of used car

x : car attributesy : price

y = g (x | q )g ( ) model, q parameters


y = wx+w0

21

Uses of Supervised Learning

Prediction of future cases: Use the rule to predict the output for future inputs

Knowledge extraction: The rule is easy to understand

Compression: The rule is simpler than the data it explains

Outlier detection: Exceptions that are not covered by the rule, e.g.,

fraud

Lecture 1: Introduction

22

Unsupervised Learning Finding regularities in data No mapping to outputs Clustering:

Grouping similar instances Example applications:

Customer segmentation in CRM Image compression: Color quantization NLP: Unsupervised text categorization


23

Reinforcement Learning Learning a policy = sequence of outputs/actions No supervised output but delayed reward Example applications:

Game playing Robot in a maze NLP: Dialogue systems, for example:

NJFun: A Reinforcement Learning Spoken Dialogue System (http://acl.ldc.upenn.edu/W/W00/W00-0304.pdf)

Reinforcement Learning for Spoken Dialogue Systems: Comparing Strengths and Weaknesses for Practical Deployment (http://research.microsoft.com/apps/pubs/default.aspx?id=70295)



Supervised Learning

Introduction to:•Margin•Noise•Bias


Supervised Classification Learning the class C of a “family car” from

examples Prediction: Is car x a family car? Knowledge extraction: What do people expect

from a family car? Output (labels):

Positive (+) and negative (–) examples Input representation (features):

x1: price, x2 : engine power

26

Training set X

X {x t,r t }t1N

r 1 if x is positive

0 if x is negative


x x1

x2


Hypothesis class H

p1 price p2 AND e1 engine power e2

28

Empirical (training) error


h(x) 1 if h says x is positive

0 if h says x is negative

E(h | X ) 1 h x t r t t1

N

Empirical error of h on X:


S, G, and the Version Space

most specific hypothesis, S

most general hypothesis, G

h Î H, between S and G is consistent [E( h | X) = 0] and make up the version space


Margin Choose h with largest margin


NoiseUnwanted anomaly in data Imprecision in input attributes Errors in labeling data points Hidden attributes (relative to H)

Consequence: No h in H may be consistent!


Noise and Model ComplexityArguments for simpler model (Occam’s razor

principle)1. Easier to make predictions

2. Easier to train (fewer parameters)

3. Easier to understand

4. Generalizes better (if data is noisy)


Inductive Bias Learning is an ill-posed problem

Training data is never sufficient to find a unique solution

There are always infinitely many consistent hypotheses

We need an inductive bias: Assumptions that entail a unique h for a training

set X1. Hypothesis class H – axis-aligned rectangles2. Learning algorithm – find consistent hypothesis with

max-margin3. Hyperparameters – trade-off between training error

and margin


Model Selection and Generalization Generalization – how well a model performs

on new data Overfitting: H more complex than C Underfitting: H less complex than C


Triple Trade-Off Trade-off between three factors:

1. Complexity of H, c(H)2. Training set size N3. Generalization error E on new data

Dependencies: As N , E¯ As c(H) , first E¯ and then E


Model Selection Generalization Error To estimate generalization error, we need data

unseen during training:

Given models (hypotheses) h1, ..., hk induced from the training set X, we can use E(hi | V ) to select the model hi with the smallest generalization error

ˆ E E(h |V ) 1 h x t r t t1

M

V {x t,r t }t1M X


Model Assessment To estimate the generalization error of the

best model hi, we need data unseen during training and model selection

Standard setup:1. Training set X (50–80%)2. Validation (development) set V (10–25%)3. Test (publication) set T (10–25%)

Note: Validation data can be added to training set before

testing Resampling methods can be used if data is limited

Cross-Validation


121

312

22

321

11

KK

KK

K

K

XXXTXV

XXXTXV

XXXTXV

K-fold cross-validation: Divide X into X1, ..., XK

Note: Generalization error estimated by means across K folds Training sets for different folds share K–2 parts Separate test set must be maintained for model

assessment

Bootstrapping


36801

1 1 .

e

N

N

Generate new training sets of size N from X by random sampling with replacement

Use original training set as validation set (V = X ) Probability that we do not pick an instance after

N draws

that is, only 36.8% of instances are new!

Measuring Error


Error rate = # of errors / # of instances = (FP+FN) / N

Accuracy = # of correct / # of instances = (TP+TN) / N

Recall = # of found positives / # of positives = TP / (TP+FN)

Precision = # of found positives / # of found = TP / (TP+FP)


Statistical Inference Interval estimation to quantify the precision of

our measurements

Hypothesis testing to assess whether differences between models are statistically significant

m 1.96N

e01 e10 1 2

e01 e10

~ X12


Supervised Learning – Summary

42

Training data + learner hypothesis Learner incorporates inductive bias

Test data + hypothesis estimated generalization Test data must be unseen

Next lectures: Different learners in LT with different inductive biases

Anatomy of a Supervised Learner (Dimensions of a supervised machine learning algorithm)

Model:

Loss function:

Optimization procedure:

g x |

E | X L r t ,g x t | t


*arg min

E | X


Reading Alpaydin (2010): Ch. 1-2; 19 (mathematical

underpinnings) Witten and Frank (2005): Ch. 1 (examples and

domains of application)


End of Lecture 1

Thanks for your attention

lecture 01: machine learning for language technology - introduction

Education