1. START OF DAY 1 Reading: Chap. 1 & 2

2. Introduction

3. How Do We Learn? By being told – Content transfer By analogy – Context transfer By induction – Knowledge construction

4. Learning by Induction Induction is a process that "involves intellectual leaps from the particular to the general" Orley’s Experience

5. Group These Objects Write down your answer – Keep it to yourself

6. UGLY PRETTY Find the Rule(s) Write down your answer – Keep it to yourself

7. Requirements for Induction A language, L p, to represent the particular (i.e., specific instances) A language, L g, to represent the general (i.e., generalizations) A matching predicate, match(G,i), that is true if G is correct about i A set, I, of particulars of some unknown general G*

8. What Does Induction Do? Observes the particulars in I Produces a generalization, G, such that: – G is consistent: For all i  I, match(G,i) – G generalizes beyond I: For many i  I, match(G,i) – G resembles G*: For most i, G(i)=G*(i)

9. Two Different Contexts Picture 1 is an example of Unsupervised Learning – There is no predefined label for the instances Picture 2 is an example of Supervised Learning – There is a predefined label (or class assignment) for the instances

10. Unsupervised Learning Consistency is (vacuously) guaranteed since you pick G (=G*) and then assign instances to groups based on G Generalization accuracy is also guaranteed, for the same reason Need a mechanism to choose among possible groupings (i.e., what makes one grouping better than another?) – Internal metrics (e.g., compactness) – External metrics (e.g., labeled data – assumes G*)

11. Supervised Learning Must try to achieve consistency – Why is that desirable? – Is it always possible? Noise (i.e., mislabeled instances) Limitation of L g (i.e., cannot represent a consistent G) Generalization accuracy can be measured directly Need a mechanism to choose among possible generalizations (i.e., what makes one generalization “better” than another? Remember, we do not know G*)

12. Revisiting Our Examples Picture 1 – What groupings did you come up with? – How good are your groupings? Example 2 – What generalizations did you come up with? – Are they consistent? – How well do you think your generalizations will do beyond the observed instances?

13. Zooming in on Picture 2 There are obviously several generalizations that are consistent: – If red or green then class 1 otherwise class 2 – If less than 3 edges then class 1 otherwise class 2 So: why did you choose the one you did? Suppose that I now tell you that the PRETTY class is the set of complex polygons (i.e., with 4 sides or more). Would that change your preference? Why? BIAS

14. The Need for and Role of Bias

15. Recall: Concept Learning Given: A language of observations/instances A language of concepts/generalizations A matching predicate A set of observations Find generalizations that: 1. Are consistent with the observations, and 2. Classify instances beyond those observed

16. Working Example (I) Observations are characterized by a fixed set of features (or attributes) and each instance corresponds to a specific assignment of values to the features For example: Features: color [red, blue, green, yellow] shape [square, triangle, circle] size [large, medium, small] Observation: O1 O2 O3, etc. Language of instances = set of attribute-value pairs Attribute-value Language (AVL)

17. Working Example (II) Generalizations are used to represent N≥1 instances For example: G1 represents all red objects (independent of color and size) G2 represents small blue objects that are either squares or triangles Each generalization can be viewed as a set, namely of the instances it represents (or matches) In the above example: G1={,,,,…, } G2={, } Language of generalizations = designer chooses

18. Our First Learning Algorithm Definitions: – T = set of training instances – S = set of maximally specific generalizations consistent with T – G = set of maximally general generalizations consistent T Version Space Algorithm: – To the smallest extent necessary to maintain consistency with T – S keeps generalizing to accommodate new positive instances – G keeps specializing to avoid new negative instances – G remains as general as possible and S remains as specific as possible

19. Version Space Learning Initialize G to the most general concept in the space Initialize S to the first positive training instance For each new positive training instance p – Delete all members of G that do not cover p – For each s in S If s does not cover p – Replace s with its most specific generalizations that cover p – Remove from S any element more general than some other element in S – Remove from S any element not more specific than some element in G For each new negative training instance n – Delete all members of S that cover n – For each g in G If g covers n – Replace g with its most general specializations that do not cover n – Remove from G any element more specific than some other element in G – Remove from G any element more specific than some element in S If G=S and both are singletons – A single concept consistent with the training data has been found If G and S become empty – There is no concept consistent with the training data

20. Visualizing the Version Space - - - - -- - -- + + ? ? ? ? Boundary of G Boundary of S

21. VS Example (I) Task: – Two-object diagrams classification Instance Language: – AVL: Size x Color x Shape Generalization Language – AVL U {*} Training Instances – P 1 : {(Large Red Triangle) (Small Blue Circle)} – P 2 : {(Large Blue Circle) (Small Red Triangle)} – N 1 : {(Large Blue Triangle) (Small Blue Triangle)}

22. VS Example (II) Initialization – G 0 = [{(* * *) (* * *)}]most general – S 0 =  most specific

23. VS Example (III) (L R T) (L R T) (S R T)(S R T) (L R T) (L R C)(L R T) (S R C)(S R T) (S R C) (L R T) (L B T)(L R T) (S B T)(S R T) (S B T) (L R T) (L B C)(L R T) (S B C)(S R T) (S B C) (L R C) (S R T) (L R C) (L R C) (S R C)(S R C) (L R C) (L B T)(L R C) (S B T)(S R C) (S B T) (L R C) (L B C)(L R C) (S B C)(S R C) (S B C) (L B T) (S R T) (L B T) (S R C) (L B T) (L B T) (S B T)(S B T) (L B T) (L B C)(L B T) (S B C)(S B T) (S B C) (L B C) (S R T) (L B C) (S R C) (L B C) (S B T) (L B C) (L B C) (S B C)(S B C)

24. VS Example (IV) After P1 – G 1 = [{(* * *) (* * *)}] no change – S 1 = [{(L R T) (S B C)}] minimum generalization to cover P 1

25. VS Example (V) (L R T) (L R T) (S R T)(S R T) (L R T) (L R C)(L R T) (S R C)(S R T) (S R C) (L R T) (L B T)(L R T) (S B T)(S R T) (S B T) (L R T) (L B C)(L R T) (S B C)(S R T) (S B C) (L R C) (S R T) (L R C) (L R C) (S R C)(S R C) (L R C) (L B T)(L R C) (S B T)(S R C) (S B T) (L R C) (L B C)(L R C) (S B C)(S R C) (S B C) (L B T) (S R T) (L B T) (S R C) (L B T) (L B T) (S B T)(S B T) (L B T) (L B C)(L B T) (S B C)(S B T) (S B C) (L B C) (S R T) (L B C) (S R C) (L B C) (S B T) (L B C) (L B C) (S B C)(S B C)

26. VS Example (VI) After P2 – G 2 = [{(* * *) (* * *)}] no change – S 2 = [{(L * *) (S * *)} {(* R T) (* B C)}] minimum generalization to cover P 2 S extends its boundary to cover the new positive instance, but only as far as needed – No zeal!

27. VS Example (VII) (L R T) (L R T) (S R T)(S R T) (L R T) (L R C)(L R T) (S R C)(S R T) (S R C) (L R T) (L B T)(L R T) (S B T)(S R T) (S B T) (L R T) (L B C)(L R T) (S B C)(S R T) (S B C) (L R C) (S R T) (L R C) (L R C) (S R C)(S R C) (L R C) (L B T)(L R C) (S B T)(S R C) (S B T) (L R C) (L B C)(L R C) (S B C)(S R C) (S B C) (L B T) (S R T) (L B T) (S R C) (L B T) (L B T) (S B T)(S B T) (L B T) (L B C)(L B T) (S B C)(S B T) (S B C) (L B C) (S R T) (L B C) (S R C) (L B C) (S B T) (L B C) (L B C) (S B C)(S B C)

28. VS Example (VIII) After N1 – G 3 = [{(* R *) (* * *)} {(* * C) (* * *)}] minimum specialization to exclude N 1 – S 3 = [{(* R T) (* B C)}] remove inconsistent generalization G contracts its boundary to exclude the new negative instance, but only as far as needed – No zeal!

29. VS Example (IX) (L R T) (L R T) (S R T)(S R T) (L R T) (L R C)(L R T) (S R C)(S R T) (S R C) (L R T) (L B T)(L R T) (S B T)(S R T) (S B T) (L R T) (L B C)(L R T) (S B C)(S R T) (S B C) (L R C) (S R T) (L R C) (L R C) (S R C)(S R C) (L R C) (L B T)(L R C) (S B T)(S R C) (S B T) (L R C) (L B C)(L R C) (S B C)(S R C) (S B C) (L B T) (S R T) (L B T) (S R C) (L B T) (L B T) (S B T)(S B T) (L B T) (L B C)(L B T) (S B C)(S B T) (S B C) (L B C) (S R T) (L B C) (S R C) (L B C) (S B T) (L B C) (L B C) (S B C)(S B C)

30. Predicting with the Version Space A new instance is classified as positive if and only if it is covered by every generalization in the version space A new instance is classified as negative if and only if no generalization in the version space covers it If some, but not all, of the generalizations in the version space cover the new instance, then the instance cannot be classified with certainty (An estimated classification based on the proportion of generalizations within the version space, that cover and do not cover the new instance, could be given)

31. VS Example (X) The Version Space is: – G 3 = [{(* R *) (* * *)} {(* * C) (* * *)}] – S 3 = [{(* R T) (* B C)}] Predicting: – {(Small Red Triangle) (Small Blue Circle)}: + – {(Small Blue Triangle) (Large Red Circle)}: - – {(Large Red Triangle) (Large Blue Triangle)}: ?

32. Taking Stock Does VS solve the Concept Learning problem? It produces generalizations The generalizations are consistent with T The generalizations extend beyond T Why/how does it work? Let’s make some assumptions and replay

33. Unbiased Generalization Language A language such that every possible subset of instances can be represented AVL U {*} is not unbiased Every replacement by * causes the representation of ALL of the values of the corresponding attribute AVL U { ∨ } is better but still not unbiased It cannot represent {, } In UGL, all subsets must have a representation I.e., UGL = power set of the given instance language

34. Unbiased Generalization Procedure Uses Unbiased Generalization Language Computes Version Space (VS) relative to UGL VS = set of all expressible generalizations consistent with the training instances (in case that was not quite clear still)

35. Claim VS with UGL cannot solve part 2 of the Concept Learning problem, i.e., learning is limited to rote learning

36. Lemma 1 Any new instance, NI, is classified as positive if and only if NI is identical to some observed positive instance Proof: (  ) If NI is identical to some observed positive instance, then NI is classified as positive – Follows directly from the definition of VS (  ) If NI is classified as positive, then NI is identical to some observed positive instance – Let g={p: p is an observed positive instance} UGL  g  VS NI matches all of VS  NI matches g

37. Lemma 2 Any new instance, NI, is classified as negative if and only if NI is identical to some observed negative instance Proof: (  ) If NI is identical to some observed negative instance, then NI is classified as negative – Follows directly from the definition of VS (  ) If NI is classified as negative, then NI is identical to some observed negative instance – Let G={all subsets containing observed negative instances} UGL  G  VS=UGL NI matches none in VS  NI was observed

38. Lemma 3 If NI is any instance which was not observed, then NI matches exactly one half of VS and cannot be classified Proof: (  ) If NI was not observed, then NI matches exactly one half of VS, and so cannot be classified – Let g={p: p is an observed positive instance} – Let G’={all subsets of unobserved instances} UGL  VS={g  g’: g’  G’} NI was not observed  NI matches exactly ½ of G’  NI matches exactly ½ of VS

39. Theorem It follows directly from Lemmas 1-3 that: An unbiased generalization procedure can never make the inductive leap necessary to classify instances beyond those it has observed

40. Another Way to Look at It… There are 2 2 n Boolean functions of n inputs x1x2x3ClassPossible Consistent Function Hypotheses 000111111111111111 11 001111111111111111 11 010111111111111111 11 011111111111111111 11 10000000000111111 11 10100001111000011 11 11000110011001100 11 111?01010101010101 01 What do we predict for 1 1 1? There are as many consistent functions predicting 0 as there are consistent functions predicting 1!

41. Yet Another Way to Look at it… If there is no bias, the outcome of the learner is highly dependent on the training data, and thus there is much variance among the models induced from different sets of observations – Learner memorizes (overfits) If there is a strong bias, the outcome of the learner is much less dependent on the training data, and thus there is little variance among induced models – Learner ignores observation Formalized as: bias-variance decomposition of error Semmelweis’ Experience

42. What is Bias? BIAS Any basis for choosing one decision over another, other than strict consistency with past observations

43. Going Back to Our Question Our example worked BECAUSE the generalization language (AVL U {*}) was not unbiased! In fact, we have just showed that: If a learning system is to be useful, it must have some form of bias

44. Humans as Learning Systems Do we have biases? – All kinds!!! Our/the representation language cannot express all possible classes of observations Our/the generalization procedure is biased – Domain knowledge (e.g., double bonds rarely break) – Intended use (e.g., ICU – relative cost) – Shared assumptions (e.g., crown, bridge – dentistry) – Simplicity and generality (e.g., white men can’t jump) – Analogy (e.g., heat vs. water flow, thin ice) – Commonsense (e.g., social interactions, pain, etc.) Survey Exercise

45. Our First Lesson The power of a generalization system follows directly from its biases Absence of bias = rote learning Progress towards understanding learning mechanisms depends upon understanding the sources of, and justification for, various biases We will consider these issues for every algorithm we will study

46. Are There Better Biases?

47. No Free Lunch Theorem A.k.a Law of Conservation for Generalization Performance (LCG) GP = Accuracy – 50% When taken across all learning tasks, the generalization performance of any learner sums to 0

48. NFL Intuition (I)

49. NFL Intuition (II)

50. NFL Intuition (III)

51. Second Lesson Whenever a learning algorithm performs well on some function, as measured by OTS generalization, it must perform poorly on some other(s) In other words, there is no universal learner, or best bias!

52. Impact on Users

53. Towards a Solution

54. Taking Stock We will study a number of learning algorithms We promise to: – Discuss their language and procedural biases – Always remember NFL

55. END OF DAY 1 Homework: Thought Questions