SlideShare a Scribd company logo
INFERENCE IN HMM AND BAYESIAN NETWORKS
DR MINAKSHI PRADEEP ATRE
PVG’S COET, PUNE
REFERENCES
 Journal paper, titled. “An Introduction of Hidden Markov Model and Bayesian Network by
Zoubain Ghahramani (2001)
 https://www.cs.ubc.ca/~murphyk/Bayes/bnintro.html
 https://www.youtube.com/watch?v=kqSzLo9fenk (Bayes Theorem and HMM)
 https://www.youtube.com/watch?v=YlL0YARYK-o (HMM)
 https://www.youtube.com/watch?v=EqUfuT3CC8s (Markov models)
 https://www.youtube.com/watch?v=5araDjcBHMQ (maths for Markov Model & time series)
 https://aimacode.github.io/aima-exercises/
 https://aimacode.github.io/aima-exercises/bayesian-learning-exercises/
 15-381: Artificial Intelligence (ppt)
WHY WE NEED PROBABILISTIC MODELS?
Probabilist
ic Models
Inferences
Unable to handle:
1) uncertain knowledge
and 2) probabilistic
reasoning
FOL
ONTOLOGY
 a set of concepts and categories in a subject area or
domain
 that shows their properties and the relations between
them
WHY WE NEED PROBABILISTIC MODELS?
 The FOL is used for inference
 There were 2 methods:
 Automated inference
 Forward checking
 Backward checking
 Resolution refutation
Propositional logic
has a very limited
ontology, making
only the
commitment that
the world consists
of facts
First order logic
overcomes the
limitation of this
ontology and has
different inference
methods
Objects
Propertie
s
Relations
Unable to handle:
uncertain
knowledge and
probabilistic
reasoning
CONTENTS : PROBABILISTIC MODELS FOR INFERENCE
 Learning and inference in hidden
Markov Model (HMM) in context with
Bayesian network
 Uncertainty and methods,
 Bayesian Probability and Belief network,
 probabilistic Reasoning,
 Generative models: Bayesian networks,
 inferences in Bayesian networks,
 Temporal models: Hidden Markov
models
Represents knowledge
and data as a fixed set
of random variables
with a joint probability
distribution
UNCERTAINTY AND METHODS
BAYESIAN PROBABILITY
 Solved examples on Bayesian Probability
SUM OF BASIC PROBABILITY FORMULAE
LIMITATIONS OF BAYESIAN
 What’s wrong with Bayesian networks
 Bayesian networks are very useful for modeling joint distributions
 • But they have their limitations:
 - Cannot account for temporal / sequence models
 - DAG’s (no self or any other loops)
CONCLUSION ( BAYESIAN LEARNING METHODS)
 Bayesian learning methods are firmly based on probability theory
and exploit advanced methods developed in statistics.
 Naïve Bayes is a simple generative model that works fairly well in
practice.
 A Bayesian network allows specifying a limited set of dependencies
using a directed graph.
 Inference algorithms allow determining the probability of values
for query variables given values for evidence variables.
WHAT ARE BELIEF NETWORKS?
 Def = efficient reasoning with probability is so new that there is one main approach—belief networks
 Conditional independence information is a vital and robust way to structure information about an
uncertain domain
 Belief networks are a natural way to represent conditional independence information
 The links between nodes represent the qualitative aspects of the domain, and the conditional probability
tables represent the quantitative aspects
 A belief network is a complete representation for the joint probability distribution for the domain, but is
often exponentially smaller in size
 Inference in belief networks means computing the probability distribution of a set of query variables,
given a set of evidence variables.
 Belief networks can reason causally, diagnostically, in mixed mode, or intercausally.
 No other uncertain reasoning mechanism can handle all these modes
 The complexity of belief network inference depends on the network structure
 In polytrees (singly connected networks), the computation time is linear in the size of the network
 There are various inference techniques for general belief networks, all of which have exponential complexity in
the worst case.
 In real domains, the local structure tends to make things more feasible, but care is needed to construct a
tractable network with more than a hundred nodes
 It is also possible to use approximation techniques, including stochastic simulation, to get an estimate of the
true probabilities with less computation
 Various alternative systems for reasoning with uncertainty have been suggested. All the truth-functional
systems have serious problems with mixed or intercausal reasoning
 In the context of using Bayes' rule, conditional independence relationships among variables can simplify
the computation of query results and greatly reduce the number of conditional probabilities that need
to be specified.
 We use a data structure called a belief network' to represent the dependence between variables and to
give a concise specification of the joint probability distribution.
 A belief network is a graph in which the following holds:
 1. A set of random variables makes up the nodes of the network.
 2. A set of directed links or arrows connects pairs of nodes. The intuitive meaning of an arrow from node X to
node Y is that X has a direct influence on Y.
 3. Each node has a conditional probability table that quantifies the effects that the parents have on the node. The
parents of a node are all those nodes that have arrows pointing to it.
 4. The graph has no directed cycles (hence is a directed, acyclic graph, or DAG).
A TYPICAL BELIEF NETWORK
 Consider the following situation.
 You have a new burglar alarm installed at home.
 It is fairly reliable at detecting a burglary, but also responds
on occasion to minor earthquakes. (This example is due to
Judea Pearl, a resident of Los Angeles; hence the acute
interest in earthquakes.)
 You also have two neighbors, John and Mary, who have
promised to call you at work when they hear the alarm.
 John always calls when he hears the alarm, but sometimes
confuses the telephone ringing with the alarm and calls
then, too.
 Mary, on the other hand, likes rather loud music and
sometimes misses the alarm altogether.
 Given the evidence of who has or has not called, we would
like to estimate the probability of a burglary.
 This simple domain is described by the belief network in
Figure
DISCUSSION (PAGE 456 RUSSELL & NORVIG)
 The topology of the network can be thought of as an abstract knowledge
base that holds in a wide variety of different settings, because it represents
the general structure of the causal processes in the domain rather than any
details of the population of individuals.
 In the case of the burglary network, the topology shows that burglary and
earthquakes directly affect the probability of the alarm going off, but whether
or not John and Mary call depends only on the alarm—the network thus
represents our assumption that they do not perceive any burglaries directly,
and they do not feel the minor earthquakes.
BELIEF NETWORKS : LEARNING IN BELIEF NETWORKS
 There are four kinds of belief networks, depending upon whether the structure of the network is known or
unknown, and whether the variables in the network are observable or hidden.
 known structure, fully observable -- In this case the only learnable part is the conditional probability tables.
These can be estimated directly using the statistics of the sample data set.
 unknown structure, fully observable -- Here the problem is to reconstruct the network topology. The
problem can be thought of as a search through structure space, and fitting data to each structure reduces to
the fixed-structure problem, so the MAP or ML probability value can be used as a heuristic in hill-climbing or
SA search.
 known structure, hidden variables -- This is analagous to neural network learning.
 unknown structure, hidden variables -- When some variables are unobservable, it becomes difficult to apply
prior techniques for recovering structure, but they require averaging over all possible values of the unknown
variables. No good general algorithms are known for handling this case.
HMM: TEMPORAL MODELS
Inference
has 3
factors
filtering
prediction
smoothing
HMM: TEMPORAL MODELS
 Inference in temporal model :
 filtering
 prediction
 smoothing
INFERENCE: MATHEMATICAL REPRESENTATION
MARKOV CHAIN
MARKOV CHAIN & EXAMPLE
WHAT’S NEXT : UNIT 4 LEARNING
LEARNING ALGORITHMS
1
Bayesian (naïve Bayes)
2
Decision tree
3
Neural Networks
THANK YOU

More Related Content

What's hot

Mining co-expression network
Mining co-expression networkMining co-expression network
Mining co-expression network
tuxette
 
Jack
JackJack
project_presentation
project_presentationproject_presentation
project_presentation
Russell Jarvis
 
Joint gene network inference with multiple samples: a bootstrapped consensual...
Joint gene network inference with multiple samples: a bootstrapped consensual...Joint gene network inference with multiple samples: a bootstrapped consensual...
Joint gene network inference with multiple samples: a bootstrapped consensual...
tuxette
 
Network analysis for computational biology
Network analysis for computational biologyNetwork analysis for computational biology
Network analysis for computational biology
tuxette
 
Artificial neural networks
Artificial neural networksArtificial neural networks
Artificial neural networks
jabedskakib
 
A SUPERPROCESS WITH UPPER CONFIDENCE BOUNDS FOR COOPERATIVE SPECTRUM SHARING
A SUPERPROCESS WITH UPPER CONFIDENCE BOUNDS  FOR COOPERATIVE SPECTRUM SHARINGA SUPERPROCESS WITH UPPER CONFIDENCE BOUNDS  FOR COOPERATIVE SPECTRUM SHARING
A SUPERPROCESS WITH UPPER CONFIDENCE BOUNDS FOR COOPERATIVE SPECTRUM SHARING
Nexgen Technology
 
Artificial neural networks and its application
Artificial neural networks and its applicationArtificial neural networks and its application
Artificial neural networks and its application
Hưng Đặng
 
Validation of Wireless and Mobile Network Models and Simulation Paper Review
Validation of Wireless and Mobile Network Models and Simulation Paper ReviewValidation of Wireless and Mobile Network Models and Simulation Paper Review
Validation of Wireless and Mobile Network Models and Simulation Paper Review
ZunAib Ali
 

What's hot (9)

Mining co-expression network
Mining co-expression networkMining co-expression network
Mining co-expression network
 
Jack
JackJack
Jack
 
project_presentation
project_presentationproject_presentation
project_presentation
 
Joint gene network inference with multiple samples: a bootstrapped consensual...
Joint gene network inference with multiple samples: a bootstrapped consensual...Joint gene network inference with multiple samples: a bootstrapped consensual...
Joint gene network inference with multiple samples: a bootstrapped consensual...
 
Network analysis for computational biology
Network analysis for computational biologyNetwork analysis for computational biology
Network analysis for computational biology
 
Artificial neural networks
Artificial neural networksArtificial neural networks
Artificial neural networks
 
A SUPERPROCESS WITH UPPER CONFIDENCE BOUNDS FOR COOPERATIVE SPECTRUM SHARING
A SUPERPROCESS WITH UPPER CONFIDENCE BOUNDS  FOR COOPERATIVE SPECTRUM SHARINGA SUPERPROCESS WITH UPPER CONFIDENCE BOUNDS  FOR COOPERATIVE SPECTRUM SHARING
A SUPERPROCESS WITH UPPER CONFIDENCE BOUNDS FOR COOPERATIVE SPECTRUM SHARING
 
Artificial neural networks and its application
Artificial neural networks and its applicationArtificial neural networks and its application
Artificial neural networks and its application
 
Validation of Wireless and Mobile Network Models and Simulation Paper Review
Validation of Wireless and Mobile Network Models and Simulation Paper ReviewValidation of Wireless and Mobile Network Models and Simulation Paper Review
Validation of Wireless and Mobile Network Models and Simulation Paper Review
 

Similar to Inference in HMM and Bayesian Models

NIPS2007: deep belief nets
NIPS2007: deep belief netsNIPS2007: deep belief nets
NIPS2007: deep belief nets
zukun
 
712201907
712201907712201907
712201907
IJRAT
 
Bayesian Networks and Association Analysis
Bayesian Networks and Association AnalysisBayesian Networks and Association Analysis
Bayesian Networks and Association Analysis
Adnan Masood
 
Introduction Of Artificial neural network
Introduction Of Artificial neural networkIntroduction Of Artificial neural network
Introduction Of Artificial neural network
Nagarajan
 
Quantum neural network
Quantum neural networkQuantum neural network
Quantum neural network
surat murthy
 
Modeling the pairwise key predistribution scheme in the presence of unreliabl...
Modeling the pairwise key predistribution scheme in the presence of unreliabl...Modeling the pairwise key predistribution scheme in the presence of unreliabl...
Modeling the pairwise key predistribution scheme in the presence of unreliabl...
JPINFOTECH JAYAPRAKASH
 
An information-theoretic, all-scales approach to comparing networks
An information-theoretic, all-scales approach to comparing networksAn information-theoretic, all-scales approach to comparing networks
An information-theoretic, all-scales approach to comparing networks
Jim Bagrow
 
Ijciet 10 01_153-2
Ijciet 10 01_153-2Ijciet 10 01_153-2
Ijciet 10 01_153-2
IAEME Publication
 
Artificial neural networks and its application
Artificial neural networks and its applicationArtificial neural networks and its application
Artificial neural networks and its application
Hưng Đặng
 
machinelearningengineeringslideshare-160909192132 (1).pdf
machinelearningengineeringslideshare-160909192132 (1).pdfmachinelearningengineeringslideshare-160909192132 (1).pdf
machinelearningengineeringslideshare-160909192132 (1).pdf
ShivareddyGangam
 
Artificial neural networks
Artificial neural networks Artificial neural networks
Artificial neural networks
ShwethaShreeS
 
Project Report -Vaibhav
Project Report -VaibhavProject Report -Vaibhav
Project Report -Vaibhav
Vaibhav Dhattarwal
 
Artificial Neural Networks ppt.pptx for final sem cse
Artificial Neural Networks  ppt.pptx for final sem cseArtificial Neural Networks  ppt.pptx for final sem cse
Artificial Neural Networks ppt.pptx for final sem cse
NaveenBhajantri1
 
Expandable bayesian
Expandable bayesianExpandable bayesian
Expandable bayesian
Ahmad Amri
 
Topology ppt
Topology pptTopology ppt
Topology ppt
Daksh Bapna
 
Applications in Machine Learning
Applications in Machine LearningApplications in Machine Learning
Applications in Machine Learning
Joel Graff
 
Topology ppt
Topology pptTopology ppt
Topology ppt
karan saini
 
Topology ppt
Topology pptTopology ppt
Topology ppt
boocse11
 
What Is a Neural Network
What Is a Neural NetworkWhat Is a Neural Network
What Is a Neural Network
April Joy Getigan
 
A SELF-ORGANIZING RECURRENT NEURAL NETWORK
A SELF-ORGANIZING RECURRENT NEURAL NETWORKA SELF-ORGANIZING RECURRENT NEURAL NETWORK
A SELF-ORGANIZING RECURRENT NEURAL NETWORK
ijaia
 

Similar to Inference in HMM and Bayesian Models (20)

NIPS2007: deep belief nets
NIPS2007: deep belief netsNIPS2007: deep belief nets
NIPS2007: deep belief nets
 
712201907
712201907712201907
712201907
 
Bayesian Networks and Association Analysis
Bayesian Networks and Association AnalysisBayesian Networks and Association Analysis
Bayesian Networks and Association Analysis
 
Introduction Of Artificial neural network
Introduction Of Artificial neural networkIntroduction Of Artificial neural network
Introduction Of Artificial neural network
 
Quantum neural network
Quantum neural networkQuantum neural network
Quantum neural network
 
Modeling the pairwise key predistribution scheme in the presence of unreliabl...
Modeling the pairwise key predistribution scheme in the presence of unreliabl...Modeling the pairwise key predistribution scheme in the presence of unreliabl...
Modeling the pairwise key predistribution scheme in the presence of unreliabl...
 
An information-theoretic, all-scales approach to comparing networks
An information-theoretic, all-scales approach to comparing networksAn information-theoretic, all-scales approach to comparing networks
An information-theoretic, all-scales approach to comparing networks
 
Ijciet 10 01_153-2
Ijciet 10 01_153-2Ijciet 10 01_153-2
Ijciet 10 01_153-2
 
Artificial neural networks and its application
Artificial neural networks and its applicationArtificial neural networks and its application
Artificial neural networks and its application
 
machinelearningengineeringslideshare-160909192132 (1).pdf
machinelearningengineeringslideshare-160909192132 (1).pdfmachinelearningengineeringslideshare-160909192132 (1).pdf
machinelearningengineeringslideshare-160909192132 (1).pdf
 
Artificial neural networks
Artificial neural networks Artificial neural networks
Artificial neural networks
 
Project Report -Vaibhav
Project Report -VaibhavProject Report -Vaibhav
Project Report -Vaibhav
 
Artificial Neural Networks ppt.pptx for final sem cse
Artificial Neural Networks  ppt.pptx for final sem cseArtificial Neural Networks  ppt.pptx for final sem cse
Artificial Neural Networks ppt.pptx for final sem cse
 
Expandable bayesian
Expandable bayesianExpandable bayesian
Expandable bayesian
 
Topology ppt
Topology pptTopology ppt
Topology ppt
 
Applications in Machine Learning
Applications in Machine LearningApplications in Machine Learning
Applications in Machine Learning
 
Topology ppt
Topology pptTopology ppt
Topology ppt
 
Topology ppt
Topology pptTopology ppt
Topology ppt
 
What Is a Neural Network
What Is a Neural NetworkWhat Is a Neural Network
What Is a Neural Network
 
A SELF-ORGANIZING RECURRENT NEURAL NETWORK
A SELF-ORGANIZING RECURRENT NEURAL NETWORKA SELF-ORGANIZING RECURRENT NEURAL NETWORK
A SELF-ORGANIZING RECURRENT NEURAL NETWORK
 

More from Minakshi Atre

Part1 speech basics
Part1 speech basicsPart1 speech basics
Part1 speech basics
Minakshi Atre
 
Signals&Systems: Quick pointers to Fundamentals
Signals&Systems: Quick pointers to FundamentalsSignals&Systems: Quick pointers to Fundamentals
Signals&Systems: Quick pointers to Fundamentals
Minakshi Atre
 
Unit 4 Statistical Learning Methods: EM algorithm
Unit 4 Statistical Learning Methods: EM algorithmUnit 4 Statistical Learning Methods: EM algorithm
Unit 4 Statistical Learning Methods: EM algorithm
Minakshi Atre
 
Artificial Intelligence: Basic Terminologies
Artificial Intelligence: Basic TerminologiesArtificial Intelligence: Basic Terminologies
Artificial Intelligence: Basic Terminologies
Minakshi Atre
 
2)local search algorithms
2)local search algorithms2)local search algorithms
2)local search algorithms
Minakshi Atre
 
Performance appraisal/ assessment in higher educational institutes (HEI)
Performance appraisal/ assessment in higher educational institutes (HEI)Performance appraisal/ assessment in higher educational institutes (HEI)
Performance appraisal/ assessment in higher educational institutes (HEI)
Minakshi Atre
 
DSP preliminaries
DSP preliminariesDSP preliminaries
DSP preliminaries
Minakshi Atre
 
Artificial intelligence agents and environment
Artificial intelligence agents and environmentArtificial intelligence agents and environment
Artificial intelligence agents and environment
Minakshi Atre
 
Unit 6: DSP applications
Unit 6: DSP applications Unit 6: DSP applications
Unit 6: DSP applications
Minakshi Atre
 
Unit 6: DSP applications
Unit 6: DSP applicationsUnit 6: DSP applications
Unit 6: DSP applications
Minakshi Atre
 
Learning occam razor
Learning occam razorLearning occam razor
Learning occam razor
Minakshi Atre
 
Learning in AI
Learning in AILearning in AI
Learning in AI
Minakshi Atre
 
Waltz algorithm in artificial intelligence
Waltz algorithm in artificial intelligenceWaltz algorithm in artificial intelligence
Waltz algorithm in artificial intelligence
Minakshi Atre
 
Perception in artificial intelligence
Perception in artificial intelligencePerception in artificial intelligence
Perception in artificial intelligence
Minakshi Atre
 
Popular search algorithms
Popular search algorithmsPopular search algorithms
Popular search algorithms
Minakshi Atre
 
Artificial Intelligence Terminologies
Artificial Intelligence TerminologiesArtificial Intelligence Terminologies
Artificial Intelligence Terminologies
Minakshi Atre
 
composite video signal
composite video signalcomposite video signal
composite video signal
Minakshi Atre
 
Basic terminologies of television
Basic terminologies of televisionBasic terminologies of television
Basic terminologies of television
Minakshi Atre
 
Mpeg 2
Mpeg 2Mpeg 2
Beginning of dtv
Beginning of dtvBeginning of dtv
Beginning of dtv
Minakshi Atre
 

More from Minakshi Atre (20)

Part1 speech basics
Part1 speech basicsPart1 speech basics
Part1 speech basics
 
Signals&Systems: Quick pointers to Fundamentals
Signals&Systems: Quick pointers to FundamentalsSignals&Systems: Quick pointers to Fundamentals
Signals&Systems: Quick pointers to Fundamentals
 
Unit 4 Statistical Learning Methods: EM algorithm
Unit 4 Statistical Learning Methods: EM algorithmUnit 4 Statistical Learning Methods: EM algorithm
Unit 4 Statistical Learning Methods: EM algorithm
 
Artificial Intelligence: Basic Terminologies
Artificial Intelligence: Basic TerminologiesArtificial Intelligence: Basic Terminologies
Artificial Intelligence: Basic Terminologies
 
2)local search algorithms
2)local search algorithms2)local search algorithms
2)local search algorithms
 
Performance appraisal/ assessment in higher educational institutes (HEI)
Performance appraisal/ assessment in higher educational institutes (HEI)Performance appraisal/ assessment in higher educational institutes (HEI)
Performance appraisal/ assessment in higher educational institutes (HEI)
 
DSP preliminaries
DSP preliminariesDSP preliminaries
DSP preliminaries
 
Artificial intelligence agents and environment
Artificial intelligence agents and environmentArtificial intelligence agents and environment
Artificial intelligence agents and environment
 
Unit 6: DSP applications
Unit 6: DSP applications Unit 6: DSP applications
Unit 6: DSP applications
 
Unit 6: DSP applications
Unit 6: DSP applicationsUnit 6: DSP applications
Unit 6: DSP applications
 
Learning occam razor
Learning occam razorLearning occam razor
Learning occam razor
 
Learning in AI
Learning in AILearning in AI
Learning in AI
 
Waltz algorithm in artificial intelligence
Waltz algorithm in artificial intelligenceWaltz algorithm in artificial intelligence
Waltz algorithm in artificial intelligence
 
Perception in artificial intelligence
Perception in artificial intelligencePerception in artificial intelligence
Perception in artificial intelligence
 
Popular search algorithms
Popular search algorithmsPopular search algorithms
Popular search algorithms
 
Artificial Intelligence Terminologies
Artificial Intelligence TerminologiesArtificial Intelligence Terminologies
Artificial Intelligence Terminologies
 
composite video signal
composite video signalcomposite video signal
composite video signal
 
Basic terminologies of television
Basic terminologies of televisionBasic terminologies of television
Basic terminologies of television
 
Mpeg 2
Mpeg 2Mpeg 2
Mpeg 2
 
Beginning of dtv
Beginning of dtvBeginning of dtv
Beginning of dtv
 

Recently uploaded

Embedded machine learning-based road conditions and driving behavior monitoring
Embedded machine learning-based road conditions and driving behavior monitoringEmbedded machine learning-based road conditions and driving behavior monitoring
Embedded machine learning-based road conditions and driving behavior monitoring
IJECEIAES
 
People as resource Grade IX.pdf minimala
People as resource Grade IX.pdf minimalaPeople as resource Grade IX.pdf minimala
People as resource Grade IX.pdf minimala
riddhimaagrawal986
 
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
Gino153088
 
Software Quality Assurance-se412-v11.ppt
Software Quality Assurance-se412-v11.pptSoftware Quality Assurance-se412-v11.ppt
Software Quality Assurance-se412-v11.ppt
TaghreedAltamimi
 
官方认证美国密歇根州立大学毕业证学位证书原版一模一样
官方认证美国密歇根州立大学毕业证学位证书原版一模一样官方认证美国密歇根州立大学毕业证学位证书原版一模一样
官方认证美国密歇根州立大学毕业证学位证书原版一模一样
171ticu
 
Engineering Drawings Lecture Detail Drawings 2014.pdf
Engineering Drawings Lecture Detail Drawings 2014.pdfEngineering Drawings Lecture Detail Drawings 2014.pdf
Engineering Drawings Lecture Detail Drawings 2014.pdf
abbyasa1014
 
Redefining brain tumor segmentation: a cutting-edge convolutional neural netw...
Redefining brain tumor segmentation: a cutting-edge convolutional neural netw...Redefining brain tumor segmentation: a cutting-edge convolutional neural netw...
Redefining brain tumor segmentation: a cutting-edge convolutional neural netw...
IJECEIAES
 
Advanced control scheme of doubly fed induction generator for wind turbine us...
Advanced control scheme of doubly fed induction generator for wind turbine us...Advanced control scheme of doubly fed induction generator for wind turbine us...
Advanced control scheme of doubly fed induction generator for wind turbine us...
IJECEIAES
 
Material for memory and display system h
Material for memory and display system hMaterial for memory and display system h
Material for memory and display system h
gowrishankartb2005
 
Certificates - Mahmoud Mohamed Moursi Ahmed
Certificates - Mahmoud Mohamed Moursi AhmedCertificates - Mahmoud Mohamed Moursi Ahmed
Certificates - Mahmoud Mohamed Moursi Ahmed
Mahmoud Morsy
 
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
Yasser Mahgoub
 
Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...
Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...
Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...
shadow0702a
 
artificial intelligence and data science contents.pptx
artificial intelligence and data science contents.pptxartificial intelligence and data science contents.pptx
artificial intelligence and data science contents.pptx
GauravCar
 
IEEE Aerospace and Electronic Systems Society as a Graduate Student Member
IEEE Aerospace and Electronic Systems Society as a Graduate Student MemberIEEE Aerospace and Electronic Systems Society as a Graduate Student Member
IEEE Aerospace and Electronic Systems Society as a Graduate Student Member
VICTOR MAESTRE RAMIREZ
 
原版制作(Humboldt毕业证书)柏林大学毕业证学位证一模一样
原版制作(Humboldt毕业证书)柏林大学毕业证学位证一模一样原版制作(Humboldt毕业证书)柏林大学毕业证学位证一模一样
原版制作(Humboldt毕业证书)柏林大学毕业证学位证一模一样
ydzowc
 
22CYT12-Unit-V-E Waste and its Management.ppt
22CYT12-Unit-V-E Waste and its Management.ppt22CYT12-Unit-V-E Waste and its Management.ppt
22CYT12-Unit-V-E Waste and its Management.ppt
KrishnaveniKrishnara1
 
Electric vehicle and photovoltaic advanced roles in enhancing the financial p...
Electric vehicle and photovoltaic advanced roles in enhancing the financial p...Electric vehicle and photovoltaic advanced roles in enhancing the financial p...
Electric vehicle and photovoltaic advanced roles in enhancing the financial p...
IJECEIAES
 
Applications of artificial Intelligence in Mechanical Engineering.pdf
Applications of artificial Intelligence in Mechanical Engineering.pdfApplications of artificial Intelligence in Mechanical Engineering.pdf
Applications of artificial Intelligence in Mechanical Engineering.pdf
Atif Razi
 
学校原版美国波士顿大学毕业证学历学位证书原版一模一样
学校原版美国波士顿大学毕业证学历学位证书原版一模一样学校原版美国波士顿大学毕业证学历学位证书原版一模一样
学校原版美国波士顿大学毕业证学历学位证书原版一模一样
171ticu
 
132/33KV substation case study Presentation
132/33KV substation case study Presentation132/33KV substation case study Presentation
132/33KV substation case study Presentation
kandramariana6
 

Recently uploaded (20)

Embedded machine learning-based road conditions and driving behavior monitoring
Embedded machine learning-based road conditions and driving behavior monitoringEmbedded machine learning-based road conditions and driving behavior monitoring
Embedded machine learning-based road conditions and driving behavior monitoring
 
People as resource Grade IX.pdf minimala
People as resource Grade IX.pdf minimalaPeople as resource Grade IX.pdf minimala
People as resource Grade IX.pdf minimala
 
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
 
Software Quality Assurance-se412-v11.ppt
Software Quality Assurance-se412-v11.pptSoftware Quality Assurance-se412-v11.ppt
Software Quality Assurance-se412-v11.ppt
 
官方认证美国密歇根州立大学毕业证学位证书原版一模一样
官方认证美国密歇根州立大学毕业证学位证书原版一模一样官方认证美国密歇根州立大学毕业证学位证书原版一模一样
官方认证美国密歇根州立大学毕业证学位证书原版一模一样
 
Engineering Drawings Lecture Detail Drawings 2014.pdf
Engineering Drawings Lecture Detail Drawings 2014.pdfEngineering Drawings Lecture Detail Drawings 2014.pdf
Engineering Drawings Lecture Detail Drawings 2014.pdf
 
Redefining brain tumor segmentation: a cutting-edge convolutional neural netw...
Redefining brain tumor segmentation: a cutting-edge convolutional neural netw...Redefining brain tumor segmentation: a cutting-edge convolutional neural netw...
Redefining brain tumor segmentation: a cutting-edge convolutional neural netw...
 
Advanced control scheme of doubly fed induction generator for wind turbine us...
Advanced control scheme of doubly fed induction generator for wind turbine us...Advanced control scheme of doubly fed induction generator for wind turbine us...
Advanced control scheme of doubly fed induction generator for wind turbine us...
 
Material for memory and display system h
Material for memory and display system hMaterial for memory and display system h
Material for memory and display system h
 
Certificates - Mahmoud Mohamed Moursi Ahmed
Certificates - Mahmoud Mohamed Moursi AhmedCertificates - Mahmoud Mohamed Moursi Ahmed
Certificates - Mahmoud Mohamed Moursi Ahmed
 
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
 
Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...
Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...
Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...
 
artificial intelligence and data science contents.pptx
artificial intelligence and data science contents.pptxartificial intelligence and data science contents.pptx
artificial intelligence and data science contents.pptx
 
IEEE Aerospace and Electronic Systems Society as a Graduate Student Member
IEEE Aerospace and Electronic Systems Society as a Graduate Student MemberIEEE Aerospace and Electronic Systems Society as a Graduate Student Member
IEEE Aerospace and Electronic Systems Society as a Graduate Student Member
 
原版制作(Humboldt毕业证书)柏林大学毕业证学位证一模一样
原版制作(Humboldt毕业证书)柏林大学毕业证学位证一模一样原版制作(Humboldt毕业证书)柏林大学毕业证学位证一模一样
原版制作(Humboldt毕业证书)柏林大学毕业证学位证一模一样
 
22CYT12-Unit-V-E Waste and its Management.ppt
22CYT12-Unit-V-E Waste and its Management.ppt22CYT12-Unit-V-E Waste and its Management.ppt
22CYT12-Unit-V-E Waste and its Management.ppt
 
Electric vehicle and photovoltaic advanced roles in enhancing the financial p...
Electric vehicle and photovoltaic advanced roles in enhancing the financial p...Electric vehicle and photovoltaic advanced roles in enhancing the financial p...
Electric vehicle and photovoltaic advanced roles in enhancing the financial p...
 
Applications of artificial Intelligence in Mechanical Engineering.pdf
Applications of artificial Intelligence in Mechanical Engineering.pdfApplications of artificial Intelligence in Mechanical Engineering.pdf
Applications of artificial Intelligence in Mechanical Engineering.pdf
 
学校原版美国波士顿大学毕业证学历学位证书原版一模一样
学校原版美国波士顿大学毕业证学历学位证书原版一模一样学校原版美国波士顿大学毕业证学历学位证书原版一模一样
学校原版美国波士顿大学毕业证学历学位证书原版一模一样
 
132/33KV substation case study Presentation
132/33KV substation case study Presentation132/33KV substation case study Presentation
132/33KV substation case study Presentation
 

Inference in HMM and Bayesian Models

  • 1. INFERENCE IN HMM AND BAYESIAN NETWORKS DR MINAKSHI PRADEEP ATRE PVG’S COET, PUNE
  • 2. REFERENCES  Journal paper, titled. “An Introduction of Hidden Markov Model and Bayesian Network by Zoubain Ghahramani (2001)  https://www.cs.ubc.ca/~murphyk/Bayes/bnintro.html  https://www.youtube.com/watch?v=kqSzLo9fenk (Bayes Theorem and HMM)  https://www.youtube.com/watch?v=YlL0YARYK-o (HMM)  https://www.youtube.com/watch?v=EqUfuT3CC8s (Markov models)  https://www.youtube.com/watch?v=5araDjcBHMQ (maths for Markov Model & time series)  https://aimacode.github.io/aima-exercises/  https://aimacode.github.io/aima-exercises/bayesian-learning-exercises/  15-381: Artificial Intelligence (ppt)
  • 3. WHY WE NEED PROBABILISTIC MODELS? Probabilist ic Models Inferences Unable to handle: 1) uncertain knowledge and 2) probabilistic reasoning FOL
  • 4. ONTOLOGY  a set of concepts and categories in a subject area or domain  that shows their properties and the relations between them
  • 5. WHY WE NEED PROBABILISTIC MODELS?  The FOL is used for inference  There were 2 methods:  Automated inference  Forward checking  Backward checking  Resolution refutation Propositional logic has a very limited ontology, making only the commitment that the world consists of facts First order logic overcomes the limitation of this ontology and has different inference methods Objects Propertie s Relations Unable to handle: uncertain knowledge and probabilistic reasoning
  • 6. CONTENTS : PROBABILISTIC MODELS FOR INFERENCE  Learning and inference in hidden Markov Model (HMM) in context with Bayesian network  Uncertainty and methods,  Bayesian Probability and Belief network,  probabilistic Reasoning,  Generative models: Bayesian networks,  inferences in Bayesian networks,  Temporal models: Hidden Markov models Represents knowledge and data as a fixed set of random variables with a joint probability distribution
  • 8. BAYESIAN PROBABILITY  Solved examples on Bayesian Probability
  • 9. SUM OF BASIC PROBABILITY FORMULAE
  • 10. LIMITATIONS OF BAYESIAN  What’s wrong with Bayesian networks  Bayesian networks are very useful for modeling joint distributions  • But they have their limitations:  - Cannot account for temporal / sequence models  - DAG’s (no self or any other loops)
  • 11. CONCLUSION ( BAYESIAN LEARNING METHODS)  Bayesian learning methods are firmly based on probability theory and exploit advanced methods developed in statistics.  Naïve Bayes is a simple generative model that works fairly well in practice.  A Bayesian network allows specifying a limited set of dependencies using a directed graph.  Inference algorithms allow determining the probability of values for query variables given values for evidence variables.
  • 12. WHAT ARE BELIEF NETWORKS?  Def = efficient reasoning with probability is so new that there is one main approach—belief networks  Conditional independence information is a vital and robust way to structure information about an uncertain domain  Belief networks are a natural way to represent conditional independence information  The links between nodes represent the qualitative aspects of the domain, and the conditional probability tables represent the quantitative aspects  A belief network is a complete representation for the joint probability distribution for the domain, but is often exponentially smaller in size  Inference in belief networks means computing the probability distribution of a set of query variables, given a set of evidence variables.
  • 13.  Belief networks can reason causally, diagnostically, in mixed mode, or intercausally.  No other uncertain reasoning mechanism can handle all these modes  The complexity of belief network inference depends on the network structure  In polytrees (singly connected networks), the computation time is linear in the size of the network  There are various inference techniques for general belief networks, all of which have exponential complexity in the worst case.  In real domains, the local structure tends to make things more feasible, but care is needed to construct a tractable network with more than a hundred nodes  It is also possible to use approximation techniques, including stochastic simulation, to get an estimate of the true probabilities with less computation  Various alternative systems for reasoning with uncertainty have been suggested. All the truth-functional systems have serious problems with mixed or intercausal reasoning
  • 14.  In the context of using Bayes' rule, conditional independence relationships among variables can simplify the computation of query results and greatly reduce the number of conditional probabilities that need to be specified.  We use a data structure called a belief network' to represent the dependence between variables and to give a concise specification of the joint probability distribution.  A belief network is a graph in which the following holds:  1. A set of random variables makes up the nodes of the network.  2. A set of directed links or arrows connects pairs of nodes. The intuitive meaning of an arrow from node X to node Y is that X has a direct influence on Y.  3. Each node has a conditional probability table that quantifies the effects that the parents have on the node. The parents of a node are all those nodes that have arrows pointing to it.  4. The graph has no directed cycles (hence is a directed, acyclic graph, or DAG).
  • 15. A TYPICAL BELIEF NETWORK  Consider the following situation.  You have a new burglar alarm installed at home.  It is fairly reliable at detecting a burglary, but also responds on occasion to minor earthquakes. (This example is due to Judea Pearl, a resident of Los Angeles; hence the acute interest in earthquakes.)  You also have two neighbors, John and Mary, who have promised to call you at work when they hear the alarm.  John always calls when he hears the alarm, but sometimes confuses the telephone ringing with the alarm and calls then, too.  Mary, on the other hand, likes rather loud music and sometimes misses the alarm altogether.  Given the evidence of who has or has not called, we would like to estimate the probability of a burglary.  This simple domain is described by the belief network in Figure
  • 16. DISCUSSION (PAGE 456 RUSSELL & NORVIG)  The topology of the network can be thought of as an abstract knowledge base that holds in a wide variety of different settings, because it represents the general structure of the causal processes in the domain rather than any details of the population of individuals.  In the case of the burglary network, the topology shows that burglary and earthquakes directly affect the probability of the alarm going off, but whether or not John and Mary call depends only on the alarm—the network thus represents our assumption that they do not perceive any burglaries directly, and they do not feel the minor earthquakes.
  • 17. BELIEF NETWORKS : LEARNING IN BELIEF NETWORKS  There are four kinds of belief networks, depending upon whether the structure of the network is known or unknown, and whether the variables in the network are observable or hidden.  known structure, fully observable -- In this case the only learnable part is the conditional probability tables. These can be estimated directly using the statistics of the sample data set.  unknown structure, fully observable -- Here the problem is to reconstruct the network topology. The problem can be thought of as a search through structure space, and fitting data to each structure reduces to the fixed-structure problem, so the MAP or ML probability value can be used as a heuristic in hill-climbing or SA search.  known structure, hidden variables -- This is analagous to neural network learning.  unknown structure, hidden variables -- When some variables are unobservable, it becomes difficult to apply prior techniques for recovering structure, but they require averaging over all possible values of the unknown variables. No good general algorithms are known for handling this case.
  • 18. HMM: TEMPORAL MODELS Inference has 3 factors filtering prediction smoothing
  • 19. HMM: TEMPORAL MODELS  Inference in temporal model :  filtering  prediction  smoothing
  • 22. MARKOV CHAIN & EXAMPLE
  • 23. WHAT’S NEXT : UNIT 4 LEARNING
  • 24. LEARNING ALGORITHMS 1 Bayesian (naïve Bayes) 2 Decision tree 3 Neural Networks