CSCE 580 Artificial Intelligence Ch.18: Learning from Observations
description
Transcript of CSCE 580 Artificial Intelligence Ch.18: Learning from Observations
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
CSCE 580Artificial IntelligenceCh.18: Learning from
ObservationsFall 2008
Marco [email protected]
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Acknowledgment• The slides are based on the textbook [AIMA] and
other sources, including other fine textbooks and the accompanying slide sets
• The other textbooks I considered are:– David Poole, Alan Mackworth, and Randy Goebel.
Computational Intelligence: A Logical Approach. Oxford, 1998
• A second edition (by Poole and Mackworth) is under development. Dr. Poole allowed us to use a draft of it in this course
– Ivan Bratko. Prolog Programming for Artificial Intelligence, Third Edition. Addison-Wesley, 2001
• The fourth edition is under development– George F. Luger. Artificial Intelligence: Structures
and Strategies for Complex Problem Solving, Sixth Edition. Addison-Welsey, 2009
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Outline• Learning agents• Inductive learning• Decision tree learning
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Learning• Learning is essential for unknown
environments,– i.e., when designer lacks omniscience
• Learning is useful as a system construction method,– i.e., expose the agent to reality rather than
trying to write it down
• Learning modifies the agent's decision mechanisms to improve performance
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Learning agents
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Learning element• Design of a learning element is affected by
– Which components of the performance element are to be learned
– What feedback is available to learn these components
– What representation is used for the components
• Type of feedback:– Supervised learning: correct answers for
each example– Unsupervised learning: correct answers not
given– Reinforcement learning: occasional rewards
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Inductive learning• Simplest form: learn a function from examples
f is the target function
An example is a pair (x, f(x))
Problem: find a hypothesis hsuch that h ≈ fgiven a training set of examples
This is a highly simplified model of real learning:– Ignores prior knowledge– Assumes examples are given
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Inductive learning method• Construct/adjust h to agree with f on training set• (h is consistent if it agrees with f on all examples)• E.g., curve fitting:
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Inductive learning method• Construct/adjust h to agree with f on training set• (h is consistent if it agrees with f on all examples)• E.g., curve fitting:
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Inductive learning method• Construct/adjust h to agree with f on training set• h is consistent if it agrees with f on all examples• E.g., curve fitting:
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Inductive learning method• Construct/adjust h to agree with f on training set• (h is consistent if it agrees with f on all examples)• E.g., curve fitting:
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Inductive learning method• Construct/adjust h to agree with f on training set• (h is consistent if it agrees with f on all examples)• E.g., curve fitting:
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Inductive learning method• Construct/adjust h to agree with f on training set• (h is consistent if it agrees with f on all examples)• E.g., curve fitting:
• Ockham’s razor: prefer the simplest hypothesis consistent with data
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Curve Fitting and Occam’s Razor
• Data collected by Galileo in1608 – ball rolling down an inclined plane, then continuing in free-fall
• Occam's razor ( suggests the simpler model is better; it has a higher prior probability
• The simpler model may have a greater posterior probability (the plausibility of the model): Occam’s razor is not only a good heuristic, but it can be shown to follow from more fundmental principles
• Jefferys, W.H. and Berger, J.O. 1992. Ockham's razor and Bayesian analysis. American Scientist 80:64-72
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Learning decision treesProblem: decide whether to wait for a table at a restaurant,
based on the following attributes:1. Alternate: is there an alternative restaurant nearby?2. Bar: is there a comfortable bar area to wait in?3. Fri/Sat: is today Friday or Saturday?4. Hungry: are we hungry?5. Patrons: number of people in the restaurant (None,
Some, Full)6. Price: price range ($, $$, $$$)7. Raining: is it raining outside?8. Reservation: have we made a reservation?9. Type: kind of restaurant (French, Italian, Thai, Burger)10. WaitEstimate: estimated waiting time (0-10, 10-30,
30-60, >60)
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Attribute-based representations
• Examples described by attribute values (Boolean, discrete, continuous)• E.g., situations where I will/won't wait for a table:
• Classification of examples is positive (T) or negative (F)•
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Decision trees• One possible representation for hypotheses• E.g., here is the “true” tree for deciding whether to wait:
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Expressiveness• Decision trees can express any function of the input attributes• E.g., for Boolean functions, truth table row → path to leaf
• Trivially, there is a consistent decision tree for any training set with one path to leaf for each example (unless f nondeterministic in x) but it probably won't generalize to new examples
• Prefer to find more compact decision trees
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Hypothesis spacesHow many distinct decision trees with n Boolean attributes?= number of Boolean functions= number of distinct truth tables with 2n rows = 22n (for each of
the 2n rows of the decision table, the function may return 0 or 1)
• E.g., with 6 Boolean attributes, there are 18,446,744,073,709,551,616 (more than 18 quintillion) trees
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Hypothesis spacesHow many distinct decision trees with n Boolean attributes?= number of Boolean functions= number of distinct truth tables with 2n rows = 22n
• E.g., with 6 Boolean attributes, there are 18,446,744,073,709,551,616 trees
How many purely conjunctive hypotheses (e.g., Hungry Rain)?• Each attribute can be in (positive), in (negative), or out
3n distinct conjunctive hypotheses• More expressive hypothesis space
– increases chance that target function can be expressed– increases number of hypotheses consistent with training
set may get worse predictions
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Decision tree learning• Aim: find a small tree consistent with the training examples• Idea: (recursively) choose "most significant" attribute as root
of (sub)tree
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Choosing an attribute• Idea: a good attribute splits the examples into subsets
that are (ideally) "all positive" or "all negative"
• Patrons? is a better choice
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Using information theory• To implement Choose-Attribute in the DTL
algorithm• Information Content (Entropy):
I(P(v1), … , P(vn)) = Σi=1 -P(vi) log2 P(vi)• For a training set containing p positive
examples and n negative examples:npn
npn
npp
npp
npn
nppI
22 loglog),(
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Information gain• A chosen attribute A divides the training set E
into subsets E1, … , Ev according to their values for A, where A has v distinct values
• Information Gain (IG) or reduction in entropy from the attribute test:
• Choose the attribute with the largest IG
v
i ii
i
ii
iii
npn
nppI
npnpAremainder
1
),()(
)(),()( Aremaindernpn
nppIAIG
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Information gain• For the training set, p = n = 6, I(6/12, 6/12) = 1 bit
• Consider the attributes Patrons and Type (and others too):
• Patrons has the highest IG of all attributes and so is chosen by the DTL algorithm as the root
bits 0)]42,
42(
124)
42,
42(
124)
21,
21(
122)
21,
21(
122[1)(
bits 0541.)]64,
62(
126)0,1(
124)1,0(
122[1)(
IIIITypeIG
IIIPatronsIG
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Example contd.• Decision tree learned from the 12 examples:
• Substantially simpler than “true” tree---a more complex hypothesis isn’t justified by small amount of data
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Performance measurement• How do we know that h ≈ f ?
1. Use theorems of computational/statistical learning theory
2. Try h on a new test set of examples(use same distribution over example space as training set)
Learning curve = % correct on test set as a function of training set size
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Summary (so far)• Learning needed for unknown environments,
lazy designers• Learning agent = performance element +
learning element• For supervised learning, the aim is to find a
simple hypothesis approximately consistent with training examples
• Decision tree learning using information gain • Learning performance = prediction accuracy
measured on test set
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Outline for Ensemble Learning and Boosting
• Ensemble Learning– Bagging– Boosting
• Reading: [AIMA-2] Sec. 18.4• This set of slides is based on
http://www.cs.uwaterloo.ca/~ppoupart/teaching/cs486-spring05/slides/Lecture21notes.pdf
• In turn, those slides follow [AIMA-2]
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Ensemble Learning• Sometimes each learning
techniqueyields a different hypothesis• But no perfect hypothesis…• Could we combine several imperfect
hypotheses into a better hypothesis?
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Ensemble Learning• Analogies:
– Elections combine voters’ choices to pick a good candidate
– Committees combine experts’ opinions to make better decisions
• Intuitions:– Individuals often make mistakes, but the
“majority” is less likely to make mistakes.– Individuals often have partial knowledge,
but a committee can pool expertise to make better decisions
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Ensemble Learning• Definition: method to select
and combine an ensemble of hypotheses into a (hopefully) better hypothesis
• Can enlarge hypothesis space– Perceptron (a simple kind
of neural network)• linear separator
– Ensemble of perceptrons• polytope
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Bagging
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Bagging• Assumptions:
– Each hi makes error with probability p– The hypotheses are independent
• Majority voting of n hypotheses:– k hypotheses make an error:– Majority makes an error:
• – With n=5, p=0.1 error( majority ) < 0.01
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Weighted Majority• In practice
– Hypotheses rarely independent– Some hypotheses make fewer errors
than others• Let’s take a weighted majority• Intuition:
– Decrease weight of correlated hypotheses
– Increase weight of good hypotheses
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Boosting• Most popular ensemble technique• Computes a weighted majority• Can “boost” a “weak learner”• Operates on a weighted training set
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Weighted Training Set• Learning with a weighted training set
– Supervised learning -> minimize training error
– Bias algorithm to learn correctly instances with high weights
• Idea: when an instance is misclassified by a hypotheses, increase its weight so that the next hypothesis is more likely to classify it correctly
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Boosting Framework
Read the figure left to right: the algorithm builds a hypothesis on a weighted set of four examples, one hypothesis per column
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
AdaBoost (Adaptive Boosting)
There are N examples.There are M “columns” (hypotheses), each of which has weight zm
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
What can we boost?• Weak learner: produces hypotheses at
least as good as random classifier.• Examples:
– Rules of thumb– Decision stumps (decision trees of
one node)– Perceptrons– Naïve Bayes models
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Boosting Paradigm• Advantages
– No need to learn a perfect hypothesis– Can boost any weak learning algorithm– Boosting is very simple to program– Good generalization
• Paradigm shift– Don’t try to learn a perfect hypothesis– Just learn simple rules of thumbs and
boost them
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Boosting Paradigm• When we already have a bunch of
hypotheses, boosting provides a principled approach to combine them
• Useful for– Sensor fusion– Combining experts
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Boosting Applications• Any supervised learning task
– Spam filtering– Speech recognition/natural language
processing– Data mining– Etc.
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Computational Learning Theory
The slides on COLT are from ftp://ftp.cs.bham.ac.uk/pub/authors/M.Kerber/Teaching/SEM2A4/l4.ps.gzand http://www.cs.bham.ac.uk/~mmk/teaching/SEM2A4/, which also has slides on version spaces
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
How many examples are needed?
This is the probability that Hεbad
contains a consistent hypothesis
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
How many examples?
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Complexity and hypothesis language
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering
Learning Decision Lists• A decision list consists of a series of tests, each of which
is a conjunction of literals. If the tests succeeds, the decision list specifies the value to be returned. Otherwise, the processing continues with the next test in the list
• Decision lists can represent any Boolean function hence are not learnable (in polynomial time)
• A k-DL is a decision list where each test is restricted to at most k literals
• K- Dl is learnable! [Rivest, 1987]