graphical models - marc deisenroth · foundations of machine learning african masters in machine...
TRANSCRIPT
![Page 1: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/1.jpg)
Foundations of Machine LearningAfrican Masters in Machine Intelligence
Graphical ModelsMarc Deisenroth
Quantum Leap AfricaAfrican Institute for MathematicalSciences, Rwanda
Department of ComputingImperial College London
October 2, 2018
![Page 2: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/2.jpg)
Reading Material
Bishop: Pattern Recognition and Machine Learning, Chapter 8
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 2
![Page 3: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/3.jpg)
Probabilistic Models
§ Quantity of interest: Joint distribution of all observed andunobserved (latent) random variables
Probabilistic model
§ Comprises information about the prior, the likelihood and theposterior
§ Joint distribution itself can be complicated
§ Does not tell us anything about structural properties of theprobabilistic model (e.g., factorization, independence)
Probabilistic graphical models
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 3
![Page 4: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/4.jpg)
Probabilistic Models
§ Quantity of interest: Joint distribution of all observed andunobserved (latent) random variables
Probabilistic model
§ Comprises information about the prior, the likelihood and theposterior
§ Joint distribution itself can be complicated
§ Does not tell us anything about structural properties of theprobabilistic model (e.g., factorization, independence)
Probabilistic graphical models
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 3
![Page 5: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/5.jpg)
Probabilistic Graphical Models
a b
c
a b
c
a b
c
Three types of probabilistic graphical models:§ Bayesian networks (directed graphical models)§ Markov random fields (undirected graphical models)§ Factor graphs
§ Nodes: (Sets of) random variables§ Edges: Probabilistic/functional relations between variablesGraph captures the way in which the joint distribution over all
random variables can be decomposed into a product of factorsdepending only on a subset of these variables
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 4
![Page 6: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/6.jpg)
Probabilistic Graphical Models
a b
c
a b
c
a b
c
Three types of probabilistic graphical models:§ Bayesian networks (directed graphical models)§ Markov random fields (undirected graphical models)§ Factor graphs
§ Nodes: (Sets of) random variables§ Edges: Probabilistic/functional relations between variables
Graph captures the way in which the joint distribution over allrandom variables can be decomposed into a product of factorsdepending only on a subset of these variables
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 4
![Page 7: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/7.jpg)
Probabilistic Graphical Models
a b
c
a b
c
a b
c
Three types of probabilistic graphical models:§ Bayesian networks (directed graphical models)§ Markov random fields (undirected graphical models)§ Factor graphs
§ Nodes: (Sets of) random variables§ Edges: Probabilistic/functional relations between variablesGraph captures the way in which the joint distribution over all
random variables can be decomposed into a product of factorsdepending only on a subset of these variables
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 4
![Page 8: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/8.jpg)
Why are they useful?
§ Simple way to visualize the structure of a probabilistic model
§ Insights into properties of the model (e.g., conditionalindependence) by inspection of the graph
§ Can be used to design/motivate new models
§ Complex computations for inference and learning can beexpressed in terms of graphical manipulations
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 5
![Page 9: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/9.jpg)
Importance of Visualization
From Kim et al. (NIPS, 2015)
From Kim et al. (NIPS, 2015)
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 6
![Page 10: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/10.jpg)
Importance of Visualization
From Kim et al. (NIPS, 2015)
From Kim et al. (NIPS, 2015)
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 6
![Page 11: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/11.jpg)
Bayesian Networks (Directed Graphical Models)
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 7
![Page 12: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/12.jpg)
Directed Graphical Models
a
e
d
b
c
§ Nodes: Random variables
§ Shaded nodes: Observed
§ Unshaded nodes: Unobserved (latent)
§ Directed arrows from a to b: Conditional distribution ppb|aq.
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 8
![Page 13: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/13.jpg)
From Joints to Graphs
Consider the joint distribution
ppa, b, cq “ ppc|a, bqppb|aqppaq
Building the corresponding graphical model:
1. Create a node for all random variables
2. For each conditional distribution, we add a directed link (arrow)to the graph from the nodes corresponding to the variables onwhich the distribution is conditioned on
a b
c
Graph layout depends on the choice of factorization
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 9
![Page 14: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/14.jpg)
From Joints to Graphs
Consider the joint distribution
ppa, b, cq “ ppc|a, bqppb|aqppaq
Building the corresponding graphical model:
1. Create a node for all random variables
2. For each conditional distribution, we add a directed link (arrow)to the graph from the nodes corresponding to the variables onwhich the distribution is conditioned on
a b
c
Graph layout depends on the choice of factorization
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 9
![Page 15: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/15.jpg)
From Joints to Graphs
Consider the joint distribution
ppa, b, cq “ ppc|a, bqppb|aqppaq
Building the corresponding graphical model:
1. Create a node for all random variables
2. For each conditional distribution, we add a directed link (arrow)to the graph from the nodes corresponding to the variables onwhich the distribution is conditioned on
a b
c
Graph layout depends on the choice of factorizationGraphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 9
![Page 16: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/16.jpg)
From Graphs to Joints
x1 x2
x3 x4
x5
§ Joint distribution is the product of a set of conditionals, one foreach node in the graph
§ Each conditional is conditioned only on the parents of thecorresponding node in the graph
ppx1, x2, x3, x4, x5q “ ppx1qppx5qppx2|x5qppx3|x1, x2qppx4|x2q
In general: ppxq “ ppx1, . . . , xKq “śK
k“1 ppxk|parentspxkqq
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 10
![Page 17: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/17.jpg)
Graphical Model for Linear Regression
x
t
0 1
−1
0
1
From PRML (Bishop, 2006)
We are given a data setpx1, y1q, . . . , pxN , yNqwhere
yi “ f pxiq ` ε, ε „ N`
0, σ2˘
with f unknown.Find a (regression) model that
explains the data
§ Consider polynomials f pxq “řM
j“0 wjxj with parametersw “ rw0, . . . , wMs
J.
§ Bayesian linear regression: Place a conjugate Gaussian prior onthe parameters: ppwq “ N
`
0, α2I˘
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 11
![Page 18: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/18.jpg)
Graphical Model for Linear Regression
x
t
0 1
−1
0
1
From PRML (Bishop, 2006)
We are given a data setpx1, y1q, . . . , pxN , yNqwhere
yi “ f pxiq ` ε, ε „ N`
0, σ2˘
with f unknown.Find a (regression) model that
explains the data
§ Consider polynomials f pxq “řM
j“0 wjxj with parametersw “ rw0, . . . , wMs
J.
§ Bayesian linear regression: Place a conjugate Gaussian prior onthe parameters: ppwq “ N
`
0, α2I˘
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 11
![Page 19: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/19.jpg)
Graphical Model for Linear Regression
x
t
0 1
−1
0
1
From PRML (Bishop, 2006)
ppy|xq “ N`
y | f pxq, σ2˘
f pxq “Mÿ
j“0
wjxj
ppwq “ N`
0, α2I˘
w
y1 yN
w
ynN
w
ynN
xn
α
σ
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 12
![Page 20: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/20.jpg)
Graphical Model for Linear Regression
x
t
0 1
−1
0
1
From PRML (Bishop, 2006)
ppy|xq “ N`
y | f pxq, σ2˘
f pxq “Mÿ
j“0
wjxj
ppwq “ N`
0, α2I˘
w
y1 yN
w
ynN
w
ynN
xn
α
σ
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 12
![Page 21: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/21.jpg)
Graphical Model for Linear Regression
x
t
0 1
−1
0
1
From PRML (Bishop, 2006)
ppy|xq “ N`
y | f pxq, σ2˘
f pxq “Mÿ
j“0
wjxj
ppwq “ N`
0, α2I˘
w
y1 yN
w
ynN
w
ynN
xn
α
σ
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 12
![Page 22: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/22.jpg)
Conditional Independence
a b
c
a KK b|c ðñ ppa|b, cq “ ppa|cqðñ ppa, b|cq “ ppa|cqppb|cq
§ (Conditional) independence allows for a factorization of the jointdistribution More efficient inference
§ Conditional independence properties of the joint distribution canbe read directly from the graph
§ No analytical manipulations required.d-separation (Pearl, 1988)
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 13
![Page 23: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/23.jpg)
D-Separation (Directed Graphs)
AC
B
Directed, acyclic graph in which A, B, Care arbitrary, non-intersecting sets ofnodes. Does A KK B|C hold?Note: C is observed if we condition on it(and the nodes in the GM are shaded)
Consider all possible paths from any node in A to any node in B.Any such path is blocked if it includes a node such that either
§ Arrows on the path meet either head-to-tail or tail-to-tail at thenode, and the node is in the set C or
§ Arrows meet head-to-head at the node and neither the node norany of its descendants is in the set C
If all paths are blocked, then A is d-separated from B by C, and thejoint distribution satisfies A KK B|C.
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 14
![Page 24: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/24.jpg)
D-Separation (Directed Graphs)
AC
B
Directed, acyclic graph in which A, B, Care arbitrary, non-intersecting sets ofnodes. Does A KK B|C hold?Note: C is observed if we condition on it(and the nodes in the GM are shaded)
Consider all possible paths from any node in A to any node in B.Any such path is blocked if it includes a node such that either
§ Arrows on the path meet either head-to-tail or tail-to-tail at thenode, and the node is in the set C or
§ Arrows meet head-to-head at the node and neither the node norany of its descendants is in the set C
If all paths are blocked, then A is d-separated from B by C, and thejoint distribution satisfies A KK B|C.
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 14
![Page 25: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/25.jpg)
D-Separation (Directed Graphs)
AC
B
Directed, acyclic graph in which A, B, Care arbitrary, non-intersecting sets ofnodes. Does A KK B|C hold?Note: C is observed if we condition on it(and the nodes in the GM are shaded)
Consider all possible paths from any node in A to any node in B.Any such path is blocked if it includes a node such that either
§ Arrows on the path meet either head-to-tail or tail-to-tail at thenode, and the node is in the set C or
§ Arrows meet head-to-head at the node and neither the node norany of its descendants is in the set C
If all paths are blocked, then A is d-separated from B by C, and thejoint distribution satisfies A KK B|C.
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 14
![Page 26: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/26.jpg)
Example
a
e
d
b
c
(a) a KK b|c?
a
e
d
b
c
(b) a KK b|d?
A path is blocked if it includes a node such that either
§ The arrows on the path meet either head-to-tail or tail-to-tail atthe node, and the node is in the set C (observed) or
§ The arrows meet head-to-head at the node, and neither the nodenor any of its descendants is in the set C (observed)
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 15
![Page 27: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/27.jpg)
Markov Random Fields (Undirected Graphical Models)
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 16
![Page 28: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/28.jpg)
Markov Random Fields
a b
c
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 17
![Page 29: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/29.jpg)
Joint Distribution
a
b
c
d
a
b
c
d
§ Express joint distribution ppx1, . . . , xnq “: ppxq as a product offunctions defined on subsets of variables that are local to thegraph
§ If xi, xj are not connected directly by a link then xi KK xj|xztxi, xju
(conditionally independent given everything else)
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 18
![Page 30: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/30.jpg)
Joint Distribution
a
b
c
d
a
b
c
d
§ Express joint distribution ppx1, . . . , xnq “: ppxq as a product offunctions defined on subsets of variables that are local to thegraph
§ If xi, xj are not connected directly by a link then xi KK xj|xztxi, xju
(conditionally independent given everything else)
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 18
![Page 31: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/31.jpg)
Factorization of the Joint Distribution
§ If xi KK xj|xztxi, xju then xi, xj never appear in a common factor inthe factorization of the joint
Joint distribution as a product of cliques (fully connectedsubgraphs)
§ Define factors in the decomposition of the joint to be functions ofthe variables in (maximum) cliques:
ppxq9ź
CψCpxCq
Example: ppa, b, c, dq9ψ1pa, bqψ2pb, c, dq
a
b
c
d
a
b
c
d
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 19
![Page 32: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/32.jpg)
Factorization of the Joint Distribution
ppxq “1Z
ź
C
ψCpxCq
§ C: maximal clique
§ xC: all variables in this clique
§ ψCpxCq: clique potential
§ Z “ř
xś
C ψCpxCq: normalization constant
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 20
![Page 33: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/33.jpg)
Clique Potentials
ppxq “1Z
ź
C
ψCpxCq
Clique potentials ψCpxCq:
§ ψCpxCq ě 0
§ Unlike directed graphs, no probabilistic interpretation necessary(e.g., marginal or conditional).
§ If we convert a directed graph into an MRF, the clique potentialsmay have a probabilistic interpretation
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 21
![Page 34: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/34.jpg)
Normalization Constant
ppxq “1Z
ź
C
ψCpxCq
§ Gives us flexibility in the definition the factorization in an MRF
§ Normalization constant (also: partition function) Z is required forparameter learning (not covered in here)
§ In a discrete model with M discrete nodes each having K states,the evaluation Z requires summing over KM states
Exponential in the size of the model
§ In a continuous model, we need to solve integralsIntractable in many cases
Major limitation of MRFs
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 22
![Page 35: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/35.jpg)
Normalization Constant
ppxq “1Z
ź
C
ψCpxCq
§ Gives us flexibility in the definition the factorization in an MRF
§ Normalization constant (also: partition function) Z is required forparameter learning (not covered in here)
§ In a discrete model with M discrete nodes each having K states,the evaluation Z requires summing over KM states
Exponential in the size of the model
§ In a continuous model, we need to solve integralsIntractable in many cases
Major limitation of MRFsGraphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 22
![Page 36: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/36.jpg)
Conditional Independence
AC
B
Two easy checks for conditional independence:
§ A KK B|C if and only if all paths from A to B pass through C.(Then, all paths are blocked)
§ Alternative: Remove all nodes in C from the graph. If there is apath from A to B then A KK B|C does not hold
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 23
![Page 37: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/37.jpg)
Potentials as Energy Functions
§ Look only at potential functions with ψCpxCq ą 0ψCpxCq “ expp´EpxCqq for some energy function E
§ Joint distribution is the product of clique potentialsTotal energy is the sum of the energies of the clique potentials
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 24
![Page 38: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/38.jpg)
Potentials as Energy Functions
§ Look only at potential functions with ψCpxCq ą 0ψCpxCq “ expp´EpxCqq for some energy function E
§ Joint distribution is the product of clique potentialsTotal energy is the sum of the energies of the clique potentials
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 24
![Page 39: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/39.jpg)
Example: Image Restoration
From PRML (Bishop, 2006)
§ Binary image, corrupted by 10% binary noise (pixel values flipwith probability 0.1).
§ Objective: Restore noise-free imagePairwise MRF that has all its variables joined in cliques of size 2
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 25
![Page 40: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/40.jpg)
Image Restoration (2)
xi
yi
§ MRF-based approach
§ Latent variables xi P t´1,`1u are the binary noise-free pixelvalues that we wish to recover
§ Observed variables yi P t´1,`1u are the noise-corrupted pixelvalues
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 26
![Page 41: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/41.jpg)
Image Restoration (2)
xi
yi
§ MRF-based approach
§ Latent variables xi P t´1,`1u are the binary noise-free pixelvalues that we wish to recover
§ Observed variables yi P t´1,`1u are the noise-corrupted pixelvalues
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 26
![Page 42: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/42.jpg)
Clique Potentials
xi
yi
Two types of clique potentials:
§ log ψxypxi, yiq “ Epxi, yiq “ ´ηxiyi , η ą 0Strong correlation between observed and latent variables
§ log ψxxpxi, xjq “ Epxi, xjq “ ´βxixj , β ą 0for neighboring pixels xi, xj
Favor similar labels for neighboring pixels (smoothness prior)
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 27
![Page 43: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/43.jpg)
Clique Potentials
xi
yi
Two types of clique potentials:
§ log ψxypxi, yiq “ Epxi, yiq “ ´ηxiyi , η ą 0Strong correlation between observed and latent variables
§ log ψxxpxi, xjq “ Epxi, xjq “ ´βxixj , β ą 0for neighboring pixels xi, xj
Favor similar labels for neighboring pixels (smoothness prior)
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 27
![Page 44: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/44.jpg)
Energy Function
Total energy:
Epx, yq “ ´ηÿ
i
xiyi
loooomoooon
latent-observed
´βÿ
ti,ju
xixj
looooomooooon
latent-latent
`γÿ
i
xi
loomoon
bias
§ Bias term places a prior on the latent pixel values, e.g., `1.
§ Joint distribution ppx, yq “ 1Z expp´Epx, yqq
§ Fix y-values to the observed ones Implicitly define ppx|yq
§ Example of an Ising model Statistical physics
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 28
![Page 45: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/45.jpg)
ICM Algorithm for Image Restoration
Noise-corrupted image, ICM, Graph-cut (From PRML (Bishop, 2006))
Iterated Conditional Modes (ICM, Kittler & Foglein, 1984)
1. Initialize all xi “ yi
2. Pick any xj: Evaluate total energyEpxzj Y t`1u, yq, Epxzj Y t´1u, yq
3. Set xj to whichever state (˘1) has the lower energy
4. RepeatLocal optimum
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 29
![Page 46: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/46.jpg)
Relation to Directed Graphs
a b
c d
c
a b
PUD
§ Directed and undirected graphs express different conditionalindependence properties
§ Left: a KK b|H, a zKK b|c has no MRF equivalent
§ Center: a zKK b|H, c KK d|aY b, a KK b|cY d has no Bayesnetequivalent
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 30
![Page 47: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/47.jpg)
Factor Graphs
Good references:
Kschischang et al.: Factor Graphs and the Sum-Product Algorithm.IEEE Transactions on Information Theory (2001)
Loeliger: An Introduction to Factor Graphs. IEEE Signal ProcessingMagazine, (2004)
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 31
![Page 48: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/48.jpg)
Factor Graphs
a b
c
§ (Un)directed graphical models express a global function ofseveral variables as a product of factors over subsets of thosevariables
§ Factor graphs make this decomposition explicit by introducingadditional nodes for the factors themselves.
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 32
![Page 49: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/49.jpg)
Factorizing the Joint
The joint distribution is a product of factors:
ppxq “ź
sfspxsq
§ x “ px1, . . . , xnq
§ xs: Subset of variables§ fs: Factor; non-negative function of the variables xs
§ Building a factor graph as a bipartite graph:§ Nodes for all random variables (same as in (un)directed graphical
models)§ Additional nodes for factors (black squares) in the joint
distribution
§ Undirected links connecting each factor node to all of the variablenodes the factor depends on
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 33
![Page 50: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/50.jpg)
Factorizing the Joint
The joint distribution is a product of factors:
ppxq “ź
sfspxsq
§ x “ px1, . . . , xnq
§ xs: Subset of variables§ fs: Factor; non-negative function of the variables xs
§ Building a factor graph as a bipartite graph:§ Nodes for all random variables (same as in (un)directed graphical
models)§ Additional nodes for factors (black squares) in the joint
distribution
§ Undirected links connecting each factor node to all of the variablenodes the factor depends on
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 33
![Page 51: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/51.jpg)
Example
x1 x2 x3
fa fb fc fd
ppxq “ fapx1, x2q fbpx1, x2q fcpx2, x3q fdpx3q
Efficient inference algorithms for factor graphs (e.g., sum-productalgorithm, see Appendix for more information)
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 34
![Page 52: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/52.jpg)
Example
x1 x2 x3
fa fb fc fd
ppxq “ fapx1, x2q fbpx1, x2q fcpx2, x3q fdpx3q
Efficient inference algorithms for factor graphs (e.g., sum-productalgorithm, see Appendix for more information)
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 34
![Page 53: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/53.jpg)
Applications of Inference in Graphical Models
§ Ranking: TrueSkill (Herbrich et al., 2007)
§ Computer vision: de-noising, segmentation, semantic labeling, ...(e.g., Sucar & Gillies, 1994; Shotton et al., 2006; Szeliski et al., 2008)
§ Coding theory: Low-density parity-check codes, turbo codes, ...(e.g., McEliece et al., 1998)
§ Linear algebra: Solve linear equation systems (Shental et al., 2008)
§ Signal processing: Iterative state estimation (e.g., Bickson et al.,
2007; Deisenroth & Mohamed, 2012)Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 35
![Page 54: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/54.jpg)
Appendix
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 36
![Page 55: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/55.jpg)
MRF Ñ Factor Graph
1. Take variable nodes from MRF
2. Create additional factor nodes corresponding to the maximalcliques xs
3. The factors fspxsq equal the clique potentials
4. Add appropriate linksNot unique
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 37
![Page 56: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/56.jpg)
Directed Graph Ñ MRF
§ Moralization:§ Add additional undirected links between all pairs of parents for
each node in the graph§ Drop arrows on original links
§ Identify (maximum) cliques
§ Initialize all clique potentials to 1
§ Take each conditional distribution factor in the directed graph,multiply it into one of the clique potentials
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 38
![Page 57: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/57.jpg)
Example: MRF Ñ Factor Graph
x1 x2
x3
x1 x2
x3
x1 x2
x3
f fafb
§ MRF with clique potential ψpx1, x2, x3q
§ Factor graph with factor f px1, x2, x3q “ ψpx1, x2, x3q
§ Factor graph with factors, such thatfapx1, x2, x3q fbpx2, x3q “ ψpx1, x2, x3q
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 39
![Page 58: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/58.jpg)
Directed Graphical Model Ñ Factor Graph
1. Take variable nodes from Bayesian network
2. Create additional factor nodes corresponding to the conditionaldistributions
3. Add appropriate linksNot unique
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 40
![Page 59: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/59.jpg)
Example: Directed Graph Ñ Factor Graph
x1 x2
x3
x1 x2
x3
x1 x2
x3
f fcfbfa
§ Directed graph with factorization ppx1qppx2qppx3|x1, x2q
§ Factor graph with factor f px1, x2, x3q “ ppx1qppx2qppx3|x1, x2q
§ Factor graph with factors fa “ ppx1q, fb “ ppx2q, fc “ ppx3|x1, x2q
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 41
![Page 60: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/60.jpg)
Removing Cycles
x1 x2
x3 x1 x2
x3
f (x1, x2, x3)x1 x2
x3
§ Local cycles in an (un)directed graph (due to links connectingparents of a node) can be removed on conversion to a factorgraph
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 42
![Page 61: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/61.jpg)
Sum-Product Algorithm for Factor Graphs
§ Factor graphs give a uniform treatment to message passing§ Two different types of messages:
§ Messages µxÑ f pxq from variable nodes to factors§ Messages µ fÑxpxq from factors to variable nodes
§ Factors transform messages into evidence for the receiving node.
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 43
![Page 62: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/62.jpg)
Variable-to-Factor Message
xm
f1
fK
µxm→fs(x)
fs
µ fK→xm(xm)
µf1→xm (x
m )
µxmÑ fspxmq “ź
lPnepxmqz fs
µ flÑxmpxmq
§ Take the product of all incoming messages along all other links§ A variable node can send a message to a factor node once it has
received messages from all other neighboring factors§ The message that a node sends to a factor is made up of the
messages that it receives from all other factors.Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 44
![Page 63: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/63.jpg)
Factor-to-Variable Message
x1
xM
xµfs→x(x)
fsfsµ xM
→fs(x
M)
µx1→fs (x
1 )
µ fsÑxpxq “ÿ
x1
¨ ¨ ¨ÿ
xM
fspx, x1, . . . , xMqź
mPnep fsqzx
µxmÑ fspxmq
§ Take the product of the incoming messages along all other linkscoming into the factor node
§ Multiply by the factor associated with that node§ Marginalize over all of the variables associated with the
incoming messagesGraphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 45
![Page 64: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/64.jpg)
Initialization
§ If the leaf node is a variable nodes, initialize the correspondingmessages to 1:
µxÑ f pxq “ 1
§ If the leaf node is a factor node, the message should be
µ fÑxpxq “ f pxq
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 46
![Page 65: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/65.jpg)
Example (1)
x1 x2 x3
x4
fa fb
fc
From PRML (Bishop, 2006)
µx1Ñ fapx1q “ 1
µ faÑx2px2q “ÿ
x1
fapx1, x2q ¨ 1
µx4Ñ fcpx4q “ 1
µ fcÑx2px2q “ÿ
x4
fcpx2, x4q ¨ 1
µx2Ñ fbpx2q “ µ faÑx2px2qµ fcÑx2px2q
µ fbÑx3px3q “ÿ
x2
fbpx2, x3qµx2Ñ fbpx2q
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 47
![Page 66: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/66.jpg)
Example (2)
x1 x2 x3
x4
fa fb
fc
From PRML (Bishop, 2006)
µx3Ñ fbpx3q “ 1
µ fbÑx2px2q “ÿ
x3
fbpx2, x3q ¨ 1
µx2Ñ fapx2q “ µ fbÑx2px2qµ fcÑx2px2q
µ faÑx1px1q “ÿ
x2
fapx1, x2qµx2Ñ fapx2q
µx2Ñ fcpx2q “ µ faÑx2px2qµ fbÑx2px2q
µ fcÑx4px4q “ÿ
x2
fcpx2, x4qµx2Ñ fcpx2q
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 48
![Page 67: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/67.jpg)
Marginals
x
f1
fK
µx←fs(x)
fsµ fK→x(x)
µf1→x (x)
For a single variable node the marginal is given as the product of allincoming messages:
ppxq “ź
fiPnepxq
µ fiÑxpxq
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 49
![Page 68: Graphical Models - Marc Deisenroth · Foundations of Machine Learning African Masters in Machine Intelligence Graphical Models Marc Deisenroth Quantum Leap Africa African Institute](https://reader033.vdocuments.us/reader033/viewer/2022050315/5f779992f13f4c7650060f94/html5/thumbnails/68.jpg)
References I
[1] D. Bickson, D. Dolev, O. Shental, P. H. Siegel, and J. K. Wolf. Linear Detection via Belief Propagation. In Proceedings of theAnnual Allerton Conference on Communication, Control, and Computing, 2007.
[2] C. M. Bishop. Pattern Recognition and Machine Learning. Information Science and Statistics. Springer-Verlag, 2006.
[3] M. P. Deisenroth and S. Mohamed. Expectation Propagation in Gaussian Process Dynamical Systems. In Advances inNeural Information Processing Systems, pages 2618–2626, 2012.
[4] R. Herbrich, T. Minka, and T. Graepel. TrueSkill(TM): A Bayesian Skill Rating System. In Advances in Neural InformationProcessing Systems, pages 569–576. MIT Press, 2007.
[5] B. Kim, J. A. Shah, and F. Doshi-Velez. Mind the Gap: A Generative Approach to Interpretable Feature Selection andExtraction. In C. Cortes, N. D. Lawrence, D. D. Lee, M. Sugiyama, and R. Garnett, editors, Advances in Neural InformationProcessing Systems, pages 2260–2268. 2015.
[6] J. Kittler and J. Foglein. Contextual Classification of Multispectral Pixel Data. IMage and Vision Computing, 2(1):13–29, 1984.
[7] F. R. Kschischang, B. J. Frey, and H.-A. Loeliger. Factor Graphs and the Sum-Product Algorithm. IEEE Transactions onInformation Theory, 47:498–519, 2001.
[8] H.-A. Loeliger. An Introduction to Factor Graphs. IEEE Signal Processing Magazine, 21(1):28–41, 2004.
[9] R. J. McEliece, D. J. C. MacKay, and J.-F. Cheng. Turbo Decoding as an Instance of Pearl’s “Belief Propagation” Algorithm.IEEE Journal on Selected Areas in Communications, 16(2):140–152, 1998.
[10] J. Pearl. Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference. Morgan Kaufmann, 1988.
[11] O. Shental, D. Bickson, J. K. W. P. H. Siegel and, and D. Dolev. Gaussian Belief Propagatio Solver for Systems of LinearEquations. In IEEE International Symposium on Information Theory, 2008.
[12] J. Shotton, J. Winn, C. Rother, and A. Criminisi. TextonBoost: Joint Appearance, Shape and Context Modeling forMulit-Class Object Recognition and Segmentation. In Proceedings of the European Conference on Computer Vision, 2006.
[13] L. E. Sucar and D. F. Gillies. Probabilistic Reasoning in High-Level Vision. Image and Vision Computing, 12(1):42–60, 1994.
[14] R. Szeliski, R. Zabih, D. Scharstein, O. Veksler, A. A. Vladimir Kolmogorov, M. Tappen, and C. Rother. A ComparativeStudy of Energy Minimization Methods for Markov Random Fields with Smoothness-based Priors. IEEE Transactions onPattern Analysis and Machine Intelligence, 30(6):1068–1080, 2008.
Graphical Models Marc Deisenroth @AIMS, Rwanda, October 2, 2018 50