Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Graphical Models: An Introduction Lise Getoor Computer Science Dept University of Maryland http://www.cs.umd.edu/~getoor Reading List for Next Lecture • Learning Probabilistic Relational Models, L. Getoor, N. Friedman, D. Koller, A. Pfeffer. Invited contribution to the book Relational Data Mining, S. Dzeroski and N. Lavrac, Eds., Springer-Verlag, 2001. http://www.cs.umd.edu/~getoor/Publications/lprm-ch.ps http://www.cs.umd.edu/class/spring2005/cmsc828g/Readings/lp rm-ch.pdf • Probabilistic Models for Relational Data, David Heckerman, Christopher Meek and Daphne Koller http://www.cs.umd.edu/projects/srl2004/Papers/heckerman.pdff tp://ftp.research.microsoft.com/pub/tr/TR-2004-30.pdf Graphical Models • e.g. Bayesian networks, Bayes nets, Belief nets, Markov networks, HMMs, Dynamic Bayes nets, etc. • Themes: – representation – reasoning – learning • Materials based on upcoming book by Nir Friedman and Daphne Koller. • Slides based on material from Nir Friedman. Probability Distributions • Let X1,…,Xp be discrete random variables • Let P be a joint distribution over X1,…,Xp If the variables are binary, then we need O(2p) parameters to describe P Can we do better? • Key idea: use properties of independence Independent Random Variables • Two variables X and Y are independent if – P(X = x|Y = y) = P(X = x) for all values x, y – That is, learning the values of Y does not change prediction of X • If X and Y are independent then – P(X,Y) = P(X|Y)P(Y) = P(X)P(Y) • In general, if X1,…,Xp are independent, then – P(X1,…,Xp)= P(X1)...P(Xp) – Requires O(n) parameters Conditional Independence • Unfortunately, most of random variables of interest are not independent of each other • A more suitable notion is that of conditional independence • Two variables X and Y are conditionally independent given Z if – P(X = x|Y = y,Z=z) = P(X = x|Z=z) for all values x,y,z – That is, learning the values of Y does not change prediction of X once we know the value of Z – notation: I ( X , Y | Z ) Example: Naïve Bayesian Model • A common model in early diagnosis: – Symptoms are conditionally independent given the disease (or fault) • Thus, if – X1,…,Xp denote whether the symptoms exhibited by the patient (headache, high-fever, etc.) and – H denotes the hypothesis about the patients health • then, P(X1,…,Xp,H) = P(H)P(X1|H)…P(Xp|H), • This naïve Bayesian model allows compact representation – It does embody strong independence assumptions Graphical Models • Graph is language for representing independencies – Directed Acyclic Graph -> Bayesian Network – Undirected Graph -> Markov Network DAGS: Markov Assumption • We now make this independence assumption more precise for directed acyclic graphs (DAGs) • Each random variable X, is independent of its nondescendents, given its parents Pa(X) • Formally, I (X, NonDesc(X) | Pa(X)) Ancestor Parent Y1 Y2 X Non-descendent Descendent Markov Assumption Example Earthquake • In this example: – – – – – I I I I I ( ( ( ( ( E, B ) B, {E, R} ) R, {A, B, C} | E ) A, R | B,E ) C, {B, E, R} | A) Radio Burglary Alarm Call I-Maps • A DAG G is an I-Map of a distribution P if the all Markov assumptions implied by G are satisfied by P (Assuming G and P both use the same set of random variables) Examples: X Y X Y Factorization • Given that G is an I-Map of P, can we simplify the representation of P? • Example: X Y • Since I(X,Y), we have that P(X|Y) = P(X) • Applying the chain rule P(X,Y) = P(X|Y) P(Y) = P(X) P(Y) • Thus, we have a simpler representation of P(X,Y) Factorization Theorem Thm: if G is an I-Map of P, then P(X1 ,..., X p ) P(X i | Pa(X i )) i Proof: P(X i | X1 ,...,X i1 ) • By chain rule: P( X1 ,..., X p ) i • wlog. X1,…,Xp is an ordering consistent with G From assumption: Pa( X i ) {X1, , X i1 } {X1, , X i1 } Pa(X i ) NonDesc(X i ) • Since G is an I-Map, I (Xi, NonDesc(Xi)| Pa(Xi)) • Hence, I( X i , {X1, , X i1 } Pa( X i ) | Pa( X i )) • We conclude, P(Xi | X1,…,Xi-1) = P(Xi | Pa(Xi) ) Factorization Example Earthquake Radio Burglary Alarm Call P(C,A,R,E,B) = P(B)P(E|B)P(R|E,B)P(A|R,B,E)P(C|A,R,B,E) versus P(C,A,R,E,B) = P(B) P(E) P(R|E) P(A|B,E) P(C|A) Consequences • We can write P in terms of “local” conditional probabilities If G is sparse, – that is, |Pa(Xi)| < k , each conditional probability can be specified compactly – e.g. for binary variables, these require O(2k) params. representation of P is compact – linear in number of variables DAGS: Summary • The Markov Independences of a DAG G – I (Xi , NonDesc(Xi) | Pai ) • G is an I-Map of a distribution P – If P satisfies the Markov independencies implied by G • if G is an I-Map of P, then P(X1 ,..., X n ) P(X i | Pai ) i Conditional Independencies • Let Markov(G) be the set of Markov Independencies implied by G • The factorization theorem shows P (X1 ,..., Xn ) P (Xi | Pai ) G is an I-Map of P i • We can also show the opposite: Thm: of P P (X1 ,..., Xn ) P (Xi | Pai ) G is an I-Map i Implied Independencies • Does a graph G imply additional independencies as a consequence of Markov(G)? • We can define a logic of independence statements • Some axioms: – I( X ; Y | Z ) I( Y; X | Z ) – I( X ; Y1, Y2 | Z ) I( X; Y1 | Z ) d-seperation • A procedure d-sep(X; Y | Z, G) that given a DAG G, and sets X, Y, and Z returns either yes or no • Goal: d-sep(X; Y | Z, G) = yes iff I(X;Y|Z) follows from Markov(G) Paths • Intuition: dependency must “flow” along paths in the graph • A path is a sequence of neighboring variables Examples: • REAB • CAER Earthquake Radio Burglary Alarm Call Paths • We want to know when a path is – active -- creates dependency between end nodes – blocked -- cannot create dependency end nodes • We want to classify situations in which paths are active. Path Blockage Three cases: Blocked – Common cause E E – – Unblocked Active R A R A Path Blockage Three cases: – Common cause Blocked E Unblocked Active E – Intermediate cause – A A C C Path Blockage Three cases: – Common cause Blocked E – Intermediate cause – Common Effect Unblocked Active E B A B C A E C B A C Path Blockage -- General Case A path is active, given evidence Z, if • Whenever we have the configuration A C B B or one of its descendents are in Z • No other nodes in the path are in Z A path is blocked, given evidence Z, if it is not active. Example – d-sep(R,B)? E R B A C Example – d-sep(R,B) = yes – d-sep(R,B|A)? E R B A C Example – d-sep(R,B) = yes – d-sep(R,B|A) = no – d-sep(R,B|E,A)? E R B A C d-Separation • X is d-separated from Y, given Z, if all paths from a node in X to a node in Y are blocked, given Z. • Checking d-separation can be done efficiently (linear time in number of edges) – Bottom-up phase: Mark all nodes whose descendents are in Z – X to Y phase: Traverse (BFS) all edges on paths from X to Y and check if they are blocked Soundness Thm: • If – G is an I-Map of P – d-sep( X; Y | Z, G ) = yes • then – P satisfies I( X; Y | Z ) Informally, • Any independence reported by d-separation is satisfied by underlying distribution Completeness Thm: • If d-sep( X; Y | Z, G ) = no • then there is a distribution P such that – G is an I-Map of P – P does not satisfy I( X; Y | Z ) Informally, • Any independence not reported by d-separation might be violated by the underlying distribution • We cannot determine this by examining the graph structure alone I-Maps revisited • The fact that G is I-Map of P might not be that useful • For example, complete DAGs – A DAG is G is complete is we cannot add an arc without creating a cycle X2 X1 X3 X4 X2 X1 X3 • These DAGs do not imply any independencies • Thus, they are I-Maps of any distribution X4 Minimal I-Maps A DAG G is a minimal I-Map of P if • G is an I-Map of P • If G’ G, then G’ is not an I-Map of P Removing any arc from G introduces (conditional) independencies that do not hold in P Minimal I-Map Example X2 X1 • If is a minimal I-Map X3 X4 • Then, these are not I-Maps: X1 X3 X1 X3 X2 X4 X2 X4 X1 X3 X1 X3 X2 X4 X2 X4 Constructing minimal I-Maps The factorization theorem suggests an algorithm • Fix an ordering X1,…,Xn • For each i, – select Pai to be a minimal subset of {X1,…,Xi-1 }, such that I(Xi ; {X1,…,Xi-1 } - Pai | Pai ) • Clearly, the resulting graph is a minimal I-Map. Non-uniqueness of minimal I-Map • Unfortunately, there may be several minimal I-Maps for the same distribution – Applying I-Map construction procedure with different orders can lead to different structures Order: C, R, A, E, B E R Original I-Map E B A C R B A C Choosing Ordering & Causality • The choice of order can have drastic impact on the complexity of minimal I-Map • Heuristic argument: construct I-Map using causal ordering among variables • Justification? – It is often reasonable to assume that graphs of causal influence should satisfy the Markov properties. P-Maps • A DAG G is P-Map (perfect map) of a distribution P if – I(X; Y | Z) if and only if d-sep(X; Y |Z, G) = yes Notes: • A P-Map captures all the independencies in the distribution • P-Maps are unique, up to DAG equivalence P-Maps • Unfortunately, some distributions do not have a PMap Bayesian Networks • A Bayesian network specifies a probability distribution via two components: – A DAG G – A collection of conditional probability distributions P(Xi|Pai) • The joint distribution P is defined by the factorization P (X1 ,..., Xn ) P (Xi | Pai ) i • Additional requirement: G is a minimal I-Map of P DAGs and BNs • DAGs as a representation of conditional independencies: – Markov independencies of a DAG – Tight correspondence between Markov(G) and the factorization defined by G – d-separation, a sound & complete procedure for computing the consequences of the independencies – Notion of minimal I-Map – P-Maps • This theory is the basis for defining Bayesian networks Undirected Graphs: Markov Networks • Alternative representation of conditional independencies • Let U be an undirected graph • Let Ni be the set of neighbors of Xi • Define Markov(U) to be the set of independencies I( Xi ; {X1,…,Xn} - Ni - {Xi } | Ni ) • U is an I-Map of P if P satisfies Markov(U) Example This graph implies that • I(A; C | B, D ) • I(B; D | A, C ) B A D C • Note: this example does not have a directed P-Map Markov Network Factorization Thm: if • P is strictly positive, that is P(x1, …, xn ) > 0 for all assignments then • U is an I-Map of P if and only if 1 • there is a factorization P( X , , X ) f (C 1 n Z i where C1, …, Ck are the maximal cliques in U Alternative form: 1 i g (Ci ) P (X1 , , Xn ) e Z i ) Relationship between Directed & Undirected Models Chain Graphs Directed Graphs Undirected Graphs CPDs • So far, we focused on how to represent independencies using DAGs • The “other” component of a Bayesian networks is the specification of the conditional probability distributions (CPDs) • Here, we’ll just discuss the simplest representation of CPDs Tabular CPDs • When the variable of interest are all discrete, the common representation is as a table: • For example P(C|A,B) can be represented by A B P(C = 0 | A, B) P(C = 1 | A, B) 0 0 1 1 0 1 0 1 0.25 0.50 0.12 0.33 0.75 0.50 0.88 0.67 Tabular CPDs Pros: • Very flexible, can capture any CPD of discrete variables • Can be easily stored and manipulated Cons: • Representation size grows exponentially with the number of parents! • Unwieldy to assess probabilities for more than few parents Continuous CPDs • When X is a continuous variables, we need to represent the density of X, given any value of its parents – Gaussian – Conditional Gaussian CPDs: Summary • Many choices for representing CPDs • Any “statistical” model of conditional distribution can be used – e.g., any regression model • Representing structure in CPDs can have implications on independencies among variables Inference in Bayesian Networks Inference • We now have compact representations of probability distributions: – Bayesian Networks – Markov Networks • Network describes a unique probability distribution P • How do we answer queries about P? • inference is name for the process of computing answers to such queries Queries: Likelihood • There are many types of queries we might ask. • Most of these involve evidence – An evidence e is an assignment of values to a set E variables in the domain – Without loss of generality E = { Xk+1, …, Xn } • Simplest query: compute probability of evidence P (e ) P (x1 ,, xk , e ) x1 xk • This is often referred to as computing the likelihood of the evidence Queries: A posteriori belief • Often we are interested in the conditional probability of a variable given the evidence P (X , e ) P (X | e ) P (e ) • This is the a posteriori belief in X, given evidence e • A related task is computing the term P(X, e) – i.e., the likelihood of e and X = x for values of X – we can recover the a posteriori belief by P (X x , e ) P (X x | e ) P (X x , e ) x A posteriori belief This query is useful in many cases: • Prediction: what is the probability of an outcome given the starting condition – Target is a descendent of the evidence • Diagnosis: what is the probability of disease/fault given symptoms – Target is an ancestor of the evidence • Note: the direction between variables does not restrict the directions of the queries – Probabilistic inference can combine evidence form all parts of the network Queries: MAP • In this query we want to find the maximum a posteriori assignment for some variable of interest (say X1,…,Xl ) • That is, x1,…,xl maximize the probability P(x1,…,xl | e) • Note that this is equivalent to maximizing P(x1,…,xl, e) Queries: MAP We can use MAP for: • Classification – find most likely label, given the evidence • Explanation – What is the most likely scenario, given the evidence Queries: MAP Cautionary note: • The MAP depends on the set of variables • Example: – MAP of X – MAP of (X, Y) x y P(x,y) 0 0 0.35 0 1 0.05 1 0 0.3 1 1 0.3 Complexity of Inference Thm: Computing P(X = x) in a Bayesian network is NP-hard Not surprising, since we can simulate Boolean gates. Hardness • Hardness does not mean we cannot solve inference – It implies that we cannot find a general procedure that works efficiently for all networks – For particular families of networks, we can have provably efficient procedures Approaches to inference • Exact inference – Inference in Simple Chains – Variable elimination – Clustering / join tree algorithms • Approximate inference – Stochastic simulation / sampling methods – Markov chain Monte Carlo methods – Mean field theory Inference in Simple Chains X1 X2 How do we compute P(X2)? P (x2 ) P (x1 , x2 ) P (x1 )P (x2 | x1 ) x1 x1 Inference in Simple Chains (cont.) X1 X2 X3 How do we compute P(X3)? P (x3 ) P (x2, x3 ) P (x2 )P (x3 | x2 ) x2 x2 • we already know how to compute P(X2)... P (x2 ) P (x1 , x2 ) P (x1 )P (x2 | x1 ) x1 x1 Inference in Simple Chains (cont.) X1 X2 X3 ... Xn How do we compute P(Xn)? • Compute P(X1), P(X2), P(X3), … • We compute each term by using the previous one P (xi 1 ) P (xi )P (xi 1 | xi ) xi Complexity: • Each step costs O(|Val(Xi)|*|Val(Xi+1)|) operations • Compare to naïve evaluation, that requires summing over joint values of n-1 variables Inference in Simple Chains (cont.) X1 X2 • Suppose that we observe the value of X2 =x2 • How do we compute P(X1|x2)? – Recall that we it suffices to compute P(X1,x2) P (x1 , x2 ) P (x2 | x1 )P (x1 ) Inference in Simple Chains (cont.) X1 X2 X3 • Suppose that we observe the value of X3 =x3 • How do we compute P(X1,x3)? P (x1 , x3 ) P (x1 )P (x3 | x1 ) • How do we compute P(x3|x1)? P (x 3 | x1 ) P (x 2 , x 3 | x1 ) P (x 2 | x1 )P (x 3 | x1 , x 2 ) x2 P (x 2 | x1 )P (x 3 | x 2 ) x2 x2 Inference in Simple Chains (cont.) X1 X2 X3 ... Xn • Suppose that we observe the value of Xn =xn • How do we compute P(X1,xn)? P (x1 ,x n) P (x1 )P (xn | x1 ) • We compute P(xn|xn-1), P(xn|xn-2), … iteratively P (x n | x i ) P (x xi 1 i 1 , xn | xi ) P (xi 1 | xi )P (xn | xi 1 ) xi Inference in Simple Chains (cont.) X1 X2 ... Xk ... Xn • Suppose that we observe the value of Xn =xn • We want to find P(Xk|xn ) • How do we compute P(Xk,xn )? P (xk ,x n) P (xk )P (xn | xk ) • We compute P(Xk ) by forward iterations • We compute P(xn | Xk ) by backward iterations Elimination in Chains • We now try to understand the simple chain example using first-order principles A B C D • Using definition of probability, we have P (e ) P (a , b, c ,d , e ) d c b a E Elimination in Chains A B C D E • By chain decomposition, we get P (e ) P (a , b, c , d , e ) d c b a d c b a P (a )P (b | a )P (c | b )P (d | c )P (e | d ) Elimination in Chains A B C E D • Rearranging terms ... P (e ) P (a )P (b | a )P (c | b )P (d | c )P (e | d ) d c b d c b a P (c | b )P (d | c )P (e | d ) P (a )P (b | a ) a Elimination in Chains X A B C E D • Now we can perform innermost summation P (e ) P (c | b )P (d | c )P (e | d ) P (a )P (b | a ) d c b d c b a P (c | b )P (d | c )P (e | d ) p (b ) • This summation, is exactly the first step in the forward iteration we describe before Elimination in Chains X X A B C D E • Rearranging and then summing again, we get P (e ) P (c | b )P (d | c )P (e | d ) p (b ) d c d c d c b P (d | c )P (e | d ) P (c | b ) p (b ) b P (d | c )P (e | d ) p (c ) Variable Elimination General idea: • Write query in the form P (Xn , e ) P (xi | pai ) xk x3 x2 i • Iteratively – Move all irrelevant terms outside of innermost sum – Perform innermost sum, getting a new term – Insert the new term into the product A More Complex Example • “Asia” network: Visit to Asia Tuberculosis Smoking Lung Cancer Abnormality in Chest X-Ray Bronchitis Dyspnea • We want to compute P(d) • Need to eliminate: v,s,x,t,l,a,b S V L T Initial factors B A X D P (v )P (s )P (t |v )P (l | s )P (b | s )P (a |t ,l )P (x | a )P (d | a , b ) S V • We want to compute P(d) • Need to eliminate: v,s,x,t,l,a,b L T Initial factors B A X D P (v )P (s )P (t |v )P (l | s )P (b | s )P (a |t ,l )P (x | a )P (d | a , b ) Eliminate: v Compute: fv (t ) P (v )P (t |v ) v fv (t )P (s )P (l | s )P (b | s )P (a |t , l )P (x | a )P (d | a , b ) Note: fv(t) = P(t) In general, result of elimination is not necessarily a probability term • We want to compute P(d) • Need to eliminate: s,x,t,l,a,b S V L T • Initial factors B A X D P (v )P (s )P (t |v )P (l | s )P (b | s )P (a |t ,l )P (x | a )P (d | a , b ) fv (t )P (s )P (l | s )P (b | s )P (a |t , l )P (x | a )P (d | a , b ) Eliminate: s Compute: fs (b,l ) P (s )P (b | s )P (l | s ) s fv (t )fs (b, l )P (a |t , l )P (x | a )P (d | a , b ) Summing on s results in a factor with two arguments fs(b,l) In general, result of elimination may be a function of several variables S V • We want to compute P(d) • Need to eliminate: x,t,l,a,b L T • Initial factors B A X D P (v )P (s )P (t |v )P (l | s )P (b | s )P (a |t ,l )P (x | a )P (d | a , b ) fv (t )P (s )P (l | s )P (b | s )P (a |t , l )P (x | a )P (d | a , b ) fv (t )fs (b, l )P (a |t , l )P (x | a )P (d | a , b ) Eliminate: x Compute: fx (a ) P (x | a ) x fv (t )fs (b, l )fx (a )P (a |t , l )P (d | a , b ) Note: fx(a) = 1 for all values of a !! S V • We want to compute P(d) • Need to eliminate: t,l,a,b L T • Initial factors B A X D P (v )P (s )P (t |v )P (l | s )P (b | s )P (a |t ,l )P (x | a )P (d | a , b ) fv (t )P (s )P (l | s )P (b | s )P (a |t , l )P (x | a )P (d | a , b ) fv (t )fs (b, l )P (a |t , l )P (x | a )P (d | a , b ) fv (t )fs (b, l )fx (a )P (a |t , l )P (d | a , b ) Eliminate: t Compute: ft (a ,l ) fv (t )P (a |t , l ) t fs (b, l )fx (a )ft (a , l )P (d | a , b ) S V • We want to compute P(d) • Need to eliminate: l,a,b L T • Initial factors B A X D P (v )P (s )P (t |v )P (l | s )P (b | s )P (a |t ,l )P (x | a )P (d | a , b ) fv (t )P (s )P (l | s )P (b | s )P (a |t , l )P (x | a )P (d | a , b ) fv (t )fs (b, l )P (a |t , l )P (x | a )P (d | a , b ) fv (t )fs (b, l )fx (a )P (a |t , l )P (d | a , b ) fs (b, l )fx (a )ft (a , l )P (d | a , b ) Eliminate: l Compute: fl (a , b ) fs (b,l )ft (a ,l ) fl (a , b )fx (a )P (d | a , b ) l • We want to compute P(d) • Need to eliminate: b S V L T • Initial factors B A X D P (v )P (s )P (t |v )P (l | s )P (b | s )P (a |t ,l )P (x | a )P (d | a , b ) fv (t )P (s )P (l | s )P (b | s )P (a |t , l )P (x | a )P (d | a , b ) fv (t )fs (b, l )P (a |t , l )P (x | a )P (d | a , b ) fv (t )fs (b, l )fx (a )P (a |t , l )P (d | a , b ) fs (b, l )fx (a )ft (a , l )P (d | a , b ) fl (a , b )fx (a )P (d | a , b ) fa (b,d ) fb (d ) Eliminate: a,b Compute: fa (b,d ) fl (a , b )fx (a ) p (d | a , b ) a fb (d ) fa (b,d ) b Variable Elimination • We now understand variable elimination as a sequence of rewriting operations • Actual computation is done in elimination step • Exactly the same computation procedure applies to Markov networks • Computation depends on order of elimination Complexity of variable elimination • Suppose in one elimination step we compute fx (y1 ,, yk ) f 'x (x , y1 ,, yk ) x m f 'x (x , y1 , , y k ) fi (x , y1,1, , y1,li ) This requires • m Val(X ) i 1 Val(Y ) i i multiplications – For each value for x, y1, …, yk, we do m multiplications • Val(X ) Val(Yi ) additions i – For each value of y1, …, yk , we do |Val(X)| additions Complexity is exponential in number of variables in the intermediate factor! Understanding Variable Elimination • We want to select “good” elimination orderings that reduce complexity • We start by attempting to understand variable elimination via the graph we are working with • This will reduce the problem of finding good ordering to graph-theoretic operation that is well-understood Undirected graph representation • At each stage of the procedure, we have an algebraic term that we need to evaluate • In general this term is of the form: P (x1 ,, xk ) fi ( Zi ) y1 yn i where Zi are sets of variables • We now plot a graph where there is undirected edge X--Y if X,Y are arguments of some factor – that is, if X,Y are in some Zi • Note: this is the Markov network that describes the probability on the variables we did not eliminate yet Chordal Graphs • elimination ordering undirected chordal graph S V L T X D L T B A S V B A X D Graph: • Maximal cliques are factors in elimination • Factors in elimination are cliques in the graph • Complexity is exponential in size of the largest clique in graph Induced Width • The size of the largest clique in the induced graph is thus an indicator for the complexity of variable elimination • This quantity is called the induced width of a graph according to the specified ordering • Finding a good ordering for a graph is equivalent to finding the minimal induced width of the graph General Networks • From graph theory: Thm: • Finding an ordering that minimizes the induced width is NP-Hard However, • There are reasonable heuristic for finding “relatively” good ordering • There are provable approximations to the best induced width • If the graph has a small induced width, there are algorithms that find it in polynomial time Elimination on Trees • Formally, for any tree, there is an elimination ordering with induced width = 1 Thm • Inference on trees is linear in number of variables PolyTrees • A polytree is a network where there is at most one path from one variable to another A Thm: • Inference in a polytree is linear in the representation size of the network – This assumes tabular CPT representation H C B E D F G Approaches to inference • Exact inference – Inference in Simple Chains – Variable elimination – Clustering / join tree algorithms • Approximate inference – Stochastic simulation / sampling methods – Markov chain Monte Carlo methods – Mean field theory Learning Bayesian Networks Learning Bayesian networks B E Data + Prior information Inducer R A C E B P(A | E,B) e b .9 .1 e b .7 .3 e b .8 .2 e b .99 .01 Known Structure -- Complete Data E, B, A <Y,N,N> <Y,Y,Y> <N,N,Y> <N,Y,Y> . . <N,Y,Y> Inducer E B P(A | E,B) e b ? ? e b ? ? e b ? ? e b ? ? B E B E A • Network structure is specified – Inducer needs to estimate parameters • Data does not contain missing values A E B P(A | E,B) e b .9 .1 e b .7 .3 e b .8 .2 e b .99 .01 Unknown Structure -- Complete Data E, B, A <Y,N,N> <Y,Y,Y> <N,N,Y> <N,Y,Y> . . <N,Y,Y> Inducer E B P(A | E,B) e b ? ? e b ? ? e b ? ? e b ? ? B E B E A • Network structure is not specified – Inducer needs to select arcs & estimate parameters • Data does not contain missing values A E B P(A | E,B) e b .9 .1 e b .7 .3 e b .8 .2 e b .99 .01 Known Structure -- Incomplete Data E, B, A <Y,N,N> <Y,?,Y> <N,N,Y> <N,Y,?> . . <?,Y,Y> E B P(A | E,B) e b ? ? e b ? ? e b ? ? e b ? ? B E Inducer B E A • Network structure is specified • Data contains missing values – We consider assignments to missing values A E B P(A | E,B) e b .9 .1 e b .7 .3 e b .8 .2 e b .99 .01 Known Structure / Complete Data • Given a network structure G – And choice of parametric family for P(Xi|Pai) • Learn parameters for network Goal • Construct a network that is “closest” to probability that generated the data Learning Parameters for a Bayesian Network • Training data has the form: B E A E [1] B [1] A[1] C [1] D E [ M ] B [ M ] A [ M ] C [ M ] C Learning Parameters for a Bayesian Network • Since we assume i.i.d. samples, likelihood function is B E A L( : D ) P (E [m], B [m], A[m],C [m] : ) m C Learning Parameters for a Bayesian Network • By definition of network, we get B E A L( : D ) P (E [m ], B [m ], A[m ], C [m ] : ) C m P (E [m ] : ) m P (B [m ] : ) P (A[m ] | B [m ], E [m ] : ) P (C [m ] | A[m ] : ) E [1] B [1] A[1] C [1] E [ M ] B [ M ] A [ M ] C [ M ] Learning Parameters for a Bayesian Network • Rewriting terms, we get B E A L( : D ) P (E [m ], B [m], A[m ], C [m ] : ) C m P (E [m ] : ) m P (B [m] : ) m P (A[m] | B [m], E [m] : ) m P (C [m] | A[m] : ) m E [1] B [1] A[1] C [1] E [ M ] B [ M ] A [ M ] C [ M ] General Bayesian Networks Generalizing for any Bayesian network: L( : D ) P (x 1 [m ], , xn [m ] : ) i.i.d. samples m P (xi [m ] | Pai [m ] : i ) m i i m Network factorization P (xi [m ] | Pai [m ] : i ) Li (i : D ) i • The likelihood decomposes according to the structure of the network. General Bayesian Networks (Cont.) Decomposition Independent Estimation Problems If the parameters for each family are not related, then they can be estimated independently of each other. From Binomial to Multinomial • For example, suppose X can have the values 1,2,…,K • We want to learn the parameters 1, 2. …, K Sufficient statistics: • N1, N2, …, NK - the number of times each outcome is observed Likelihood function: K L( : D ) k Nk MLE: k 1 Nk ˆ k N Likelihood for Multinomial Networks • When we assume that P(Xi | Pai ) is multinomial, we get further decomposition: Li (i : D ) P (xi [m ] | Pai [m ] : i ) m P (x [m] | pa paii m ,Paii [ m ] paii i i : i ) P (xi | pai : i )N ( xii , paii ) paii xii pai xi xi |pai N ( xi , pai ) Likelihood for Multinomial Networks • When we assume that P(Xi | Pai ) is multinomial, we get further decomposition: Li (i : D ) xi |pai pai N ( xi , pai ) xi • For each value pai of the parents of Xi we get an independent multinomial problem • The MLE is N (xi , pai ) ˆ x |pa i i N ( pai ) Bayesian Approach: Dirichlet Priors • Recall that the likelihood function is K L ( : D ) k Nk k 1 • A Dirichlet prior with hyperparameters 1,…,K is defined as K P () k k 1 for legal 1,…, K k 1 Then the posterior has the same form, with hyperparameters 1+N 1,…,K +N K K K K k 1 k 1 k 1 P ( | D ) P ()P (D | ) k k 1 k Nk k k Nk 1 Dirichlet Priors (cont.) • We can compute the prediction on a new event in closed form: • If P() is Dirichlet with hyperparameters 1,…,K then P( X[1] k ) k P()d k • Since the posterior is also Dirichlet, we get P( X[M 1] k | D) k P( | D)d k Nk ( N ) Prior Knowledge • The hyperparameters 1,…,K can be thought of as “imaginary” counts from our prior experience • Equivalent sample size = 1+…+K • The larger the equivalent sample size the more confident we are in our prior Bayesian Prediction(cont.) • Given these observations, we can compute the posterior for each multinomial Xi | pai independently – The posterior is Dirichlet with parameters (Xi=1|pai)+N (Xi=1|pai),…, (Xi=k|pai)+N (Xi=k|pai) • The predictive distribution is then represented by the parameters (x i , pai ) N(x i , pai ) ~ x i|pai (pai ) N(pai ) Learning Parameters: Summary • Estimation relies on sufficient statistics – For multinomial these are of the form N (xi,pai) – Parameter estimation N (xi , pai ) ˆ x |pa i i N ( pai ) MLE (xi , pai ) N (xi , pai ) ~ x |pa i i ( pai ) N ( pai ) Bayesian (Dirichlet) • Bayesian methods also require choice of priors • Both MLE and Bayesian are asymptotically equivalent and consistent • Both can be implemented in an on-line manner by accumulating sufficient statistics Learning Structure from Complete Data Why Struggle for Accurate Structure? Earthquake Alarm Set Burglary Sound Adding an arc Earthquake Alarm Set Missing an arc Sound • • Earthquake Burglary Increases the number of parameters to be fitted Wrong assumptions about causality and domain structure Alarm Set Burglary Sound • • Cannot be compensated by accurate fitting of parameters Also misses causality and domain structure Approaches to Learning Structure • Constraint based – Perform tests of conditional independence – Search for a network that is consistent with the observed dependencies and independencies • Pros & Cons Intuitive, follows closely the construction of BNs Separates structure learning from the form of the independence tests Sensitive to errors in individual tests Approaches to Learning Structure • Score based – Define a score that evaluates how well the (in)dependencies in a structure match the observations – Search for a structure that maximizes the score • Pros & Cons Statistically motivated Can make compromises Takes the structure of conditional probabilities into account Computationally hard Likelihood Score for Structures First cut approach: – Use likelihood function • Recall, the likelihood score for a network structure and parameters is L(G, G : D) P(x1 [m], , x n [m] : G, G ) m P(x i [m] | PaiG [m] : G, G,i ) m i • Since we know how to maximize parameters from now we assume L(G : D) max G L(G , G : D) Avoiding Overfitting “Classic” issue in learning. Approaches: • Restricting the hypotheses space – Limits the overfitting capability of the learner – Example: restrict # of parents or # of parameters • Minimum description length – Description length measures complexity – Prefer models that compactly describes the training data • Bayesian methods – Average over all possible parameter values – Use prior knowledge Bayesian Inference • Bayesian Reasoning---compute expectation over unknown G P( x[M 1] | D) P(x[M 1] | D, G)P(G | D) G • Assumption: Gs are mutually exclusive and exhaustive • We know how to compute P(x[M+1]|G,D) – Same as prediction with fixed structure • How do we compute P(G|D)? Posterior Score Using Bayes rule: Prior over structures Marginal likelihood P (G | D ) P (D | G )P (G ) P (D ) Probability of Data P(D) is the same for all structures G Can be ignored when comparing structures Marginal Likelihood • By introduction of variables, we have that P (D | G ) P (D | G , )P ( | G ) d Likelihood Prior over parameters • This integral measures sensitivity to choice of parameters Marginal Likelihood for General Network The marginal likelihood has the form: P (D | G ) i paiG ( paiG ) ( paiG ) N ( paiG ) xi ( (xi , paiG ) N (xi , paiG )) ( (xi , paiG )) Dirichlet Marginal Likelihood For the sequence of values of Xi when Xi’s parents have a particular value where • N(..) are the counts from the data • (..) are the hyperparameters for each family given G Priors • We need: prior counts (..) for each network structure G • This can be a formidable task – There are exponentially many structures… BDe Score Possible solution: The BDe prior • Represent prior using two elements M0, B0 – M0 - equivalent sample size – B0 - network representing the prior probability of events BDe Score Intuition: M0 prior examples distributed by B0 • Set (xi,paiG) = M0 P(xi,paiG| B0) – Note that paiG are not the same as the parents of Xi in B0. – Compute P(xi,paiG| B0) using standard inference procedures • Such priors have desirable theoretical properties – Equivalent networks are assigned the same score Bayesian Score: Asymptotic Behavior Theorem: If the prior P( |G) is “well-behaved”, then log M log P (D | G ) l (G : D ) dim( G ) O (1) 2 Asymptotic Behavior: Consequences log M log P (D | G ) l (G : D ) dim( G ) O (1) 2 • Bayesian score is consistent – As M the “true” structure G* maximizes the score (almost surely) – For sufficiently large M, the maximal scoring structures are equivalent to G* • Observed data eventually overrides prior information – Assuming that the prior assigns positive probability to all cases Asymptotic Behavior log M Score(G : D ) l (G : D ) dim( G ) 2 • This score can also be justified by the Minimal Description Length (MDL) principle • This equation explicitly shows the tradeoff between – Fitness to data --- likelihood term – Penalty for complexity --- regularization term Scores -- Summary • Likelihood, MDL, (log) BDe have the form Score (G : D ) Score (Xi | Pa iG : N (Xi Pai )) i • BDe requires assessing prior network. It can naturally incorporate prior knowledge and previous experience • BDe is consistent and asymptotically equivalent (up to a constant) to MDL • All are score-equivalent – G equivalent to G’ Score(G) = Score(G’) Optimization Problem Input: – Training data – Scoring function (including priors, if needed) – Set of possible structures • Including prior knowledge about structure Output: – A network (or networks) that maximize the score Key Property: – Decomposability: the score of a network is a sum of terms. Heuristic Search We address the problem by using heuristic search • Define a search space: – nodes are possible structures – edges denote adjacency of structures • Traverse this space looking for high-scoring structures Search techniques: – – – – Greedy hill-climbing Best first search Simulated Annealing ... Heuristic Search (cont.) • Typical operations: S C E S C D E D S C S C E E D D Exploiting Decomposability in Local Search S S C C E E D D S C S C E E D D • Caching: To update the score of after a local change, we only need to re-score the families that were changed in the last move Greedy Hill-Climbing • Simplest heuristic local search – Start with a given network • empty network • best tree • a random network – At each iteration • Evaluate all possible changes • Apply change that leads to best improvement in score • Reiterate – Stop when no modification improves score • Each step requires evaluating approximately n new changes Greedy Hill-Climbing: Possible Pitfalls • Greedy Hill-Climbing can get struck in: – Local Maxima: • All one-edge changes reduce the score – Plateaus: • Some one-edge changes leave the score unchanged • Happens because equivalent networks received the same score and are neighbors in the search space • Both occur during structure search • Standard heuristics can escape both – Random restarts – TABU search Search: Summary • Discrete optimization problem • In general, NP-Hard – Need to resort to heuristic search – In practice, search is relatively fast (~100 vars in ~10 min): • Decomposability • Sufficient statistics • In some cases, we can reduce the search problem to an easy optimization problem – Example: learning trees Graphical Models Intro Summary • Representations – Graphs are cool way to put constraints on distributions, so that you can say lots of stuff without even looking at the numbers! • Inference – GM let you compute all kinds of different probabilities efficiently • Learning – You can even learn them auto-magically!