SlideShare a Scribd company logo
1 of 18
Download to read offline
SocNL:	
  Bayesian	
  Label	
  
Propaga5on	
  with	
  Confidence	
  
Yuto	
  Yamaguchi†	
  
Christos	
  Faloutsos‡	
  
Hiroyuki	
  Kitagawa†	
  
	
  
†U.	
  of	
  Tsukuba	
  	
  	
  	
  ‡CMU	
  
Node	
  Classifica5on	
15/01/29	
 Yuto	
  Yamaguchi	
  -­‐	
  AAAI2015	
 2	
Find: correct labels of unlabeled nodes	
?
?
Our	
  focus	
  –	
  Classifica5on	
  confidence	
  
Example	
  input	
  graph	
  
Our	
  intui5on	
  
	
  -­‐	
  A	
  is	
  most	
  probably	
  conserva5ve	
  
	
  -­‐	
  B	
  may	
  be	
  conserva5ve	
  
person	
  
è It’s	
  good	
  to	
  have	
  confidence	
  for	
  our	
  predic5on	
  
	
  
	
   	
  e.g.,	
  A	
  is	
  conserva5ve	
  with	
  confidence	
  score	
  0.9	
  
	
   	
   	
  	
  	
  B	
  is	
  conserva5ve	
  with	
  confidence	
  score	
  0.55	
  
Contribu5ons	
  
•  Novel	
  Algorithm	
  
–  Simple,	
  fast,	
  and	
  incorporates	
  confidence	
  
•  Theore5cal	
  Analysis	
  
–  Convergence	
  guarantee	
  &	
  speed	
  
–  Equivalence	
  to	
  LP	
  and	
  Bayesian	
  inference	
  
•  Empirical	
  Analysis	
  
–  Higher	
  accuracy	
  than	
  compe5tors	
  
–  Three	
  different	
  real	
  network	
  datasets	
  
PROPOSED	
  METHOD	
  
Smoothness	
  assump5on	
  
(widely	
  adopted)	
  
	
  Connected	
  nodes	
  are	
  likely	
  to	
  share	
  a	
  label	
  
B
A
Our	
  Idea	
  
B
A
Smoothness	
  assump5on	
  +	
  confidence	
  
	
  IF	
  a	
  node	
  has	
  a	
  lot	
  of	
  red/blue	
  neighbors	
  
	
  THEN	
  we	
  can	
  confidently	
  say	
  that	
  it	
  is	
  red/blue	
  
Confident	
  
Not	
  confident	
  
More	
  evidence,	
  more	
  confidence	
  	
  à	
  	
  Bayesian	
  inference	
  
Cases	
  to	
  consider	
  
•  Case1:	
  without	
  unlabeled	
  neighbors	
  
– Easy	
  but	
  unrealis5c	
  
•  Case2:	
  with	
  unlabeled	
  neighbors	
  
– We	
  want	
  to	
  handle	
  this	
  case	
  
?
?	
   ?	
  
Case1:	
  No	
  unlabeled	
  neighbors	
  
?
Prior	
  
knowledge	
  
evidence	
  
+	
  
Result	
  
Detail	
  
DCM	
  (Dirichlet	
  compound	
  mul5nomial)	
  leads	
  to	
  simple	
  result:	
  
∝	
  
fik:	
  probability	
  that	
  i	
  has	
  label	
  k	
  
nik:	
  number	
  of	
  i’s	
  neighbors	
  with	
  label	
  k	
  
αk:	
  prior	
  	
  
Case2:	
  With	
  unlabeled	
  neighbors	
  
A	
   B	
  
Classifica>on	
  result	
  for	
  A	
  affects	
  B	
  
	
  
	
  
Classifica>on	
  result	
  for	
  B	
  affects	
  A	
  
In	
  this	
  case	
  we	
  need	
  to	
  solve	
  the	
  recursive	
  equa5on:	
  
Aij:	
  entry	
  of	
  adjacency	
  matrix	
  
Detail	
  
Yes,	
  we	
  can	
  solve	
  it	
  
(Please	
  see	
  the	
  paper	
  for	
  detail)	
  
•  Simple:	
  We	
  just	
  need	
  to	
  do	
  matrix	
  inversion	
  
•  Fast:	
  power	
  itera5on	
  for	
  sparse	
  matrix	
  inversion	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  is	
  fast	
  (PUU	
  is	
  sparse)	
  
•  Confidence:	
  this	
  equa5on	
  is	
  from	
  Bayesian	
  inference	
  
THEORETICAL	
  RESULTS	
  
Convergence	
  guarantee	
  &	
  speed	
  
Theorem	
  1:	
  
The	
  itera5ve	
  algorithm	
  of	
  SocNL	
  always	
  converges	
  on	
  arbitrary	
  
graphs	
  if	
  use	
  non-­‐zero	
  prior	
  values	
  
Theorem	
  2:	
  
SocNL	
  converges	
  faster	
  if	
  use	
  larger	
  prior	
  values	
  	
  
Theorem	
  3:	
  
Time	
  complexity	
  of	
  each	
  itera5on	
  of	
  SocNL	
  is	
  O(	
  K(N+M)	
  )	
  
Equivalence	
  
Theorem	
  4:	
  
SocNL	
  is	
  equivalent	
  to	
  normal	
  LP	
  
if	
  uses	
  prior	
  values	
  =	
  0	
  
Theorem	
  5:	
  
SocNL	
  is	
  equivalent	
  to	
  Bayesian	
  inference	
  over	
  DCM	
  
if	
  ignores	
  unlabeled	
  nodes	
  
*	
  DCM:	
  Dirichlet	
  compound	
  mul5nomial	
  
EMPIRICAL	
  RESULTS	
  
Experimental	
  seings	
  
○	
  Datasets	
  
○	
  Compe5tors	
  
•  Label	
  Propaga>on	
  [ICML03]	
  
•  Myopic:	
  SocNL	
  ignoring	
  unlabeled	
  nodes	
  
Results	
  
Myopic	
  not	
  good	
  L	
  
SocNL	
  shows	
  higher	
  overall	
  accuracy	
  
than	
  compe>tors	
  	
  J	
  
Upper	
  is	
  be3er	
  	
  
Myopic	
  not	
  good	
  L	
   LP	
  not	
  good	
  L	
  
Summary	
  
•  Proposed	
  SocNL	
  
–  Simple,	
  fast,	
  and	
  incorporates	
  confidence	
  
•  Theore5cal	
  Analysis	
  
–  Convergence	
  (Theorems	
  1,2,3)	
  
–  Equivalence	
  (Theorems	
  4,5)	
  
•  Empirical	
  Analysis	
  
–  Higher	
  overall	
  accuracy	
  than	
  compe5tors	
  
Upper	
  is	
  be3er	
  	
  

More Related Content

What's hot

Mca 4040 analysis and design of algorithm
Mca 4040  analysis and design of algorithmMca 4040  analysis and design of algorithm
Mca 4040 analysis and design of algorithmsmumbahelp
 
CS 6402 – DESIGN AND ANALYSIS OF ALGORITHMS questio
CS 6402 – DESIGN AND ANALYSIS OF ALGORITHMS questioCS 6402 – DESIGN AND ANALYSIS OF ALGORITHMS questio
CS 6402 – DESIGN AND ANALYSIS OF ALGORITHMS questioKarthik Venkatachalam
 
Efficient Sparse Coding Algorithms
Efficient Sparse Coding AlgorithmsEfficient Sparse Coding Algorithms
Efficient Sparse Coding AlgorithmsAnshu Dipit
 
ICDE-2015 Shortest Path Traversal Optimization and Analysis for Large Graph C...
ICDE-2015 Shortest Path Traversal Optimization and Analysis for Large Graph C...ICDE-2015 Shortest Path Traversal Optimization and Analysis for Large Graph C...
ICDE-2015 Shortest Path Traversal Optimization and Analysis for Large Graph C...Waqas Nawaz
 
Computability and Complexity
Computability and ComplexityComputability and Complexity
Computability and ComplexityEdward Blurock
 
Divide and conquer strategy
Divide and conquer strategyDivide and conquer strategy
Divide and conquer strategyNisha Soms
 
Dynamic programmng2
Dynamic programmng2Dynamic programmng2
Dynamic programmng2debolina13
 
How Unstable is an Unstable System
How Unstable is an Unstable SystemHow Unstable is an Unstable System
How Unstable is an Unstable Systemidescitation
 
Backtracking based integer factorisation, primality testing and square root c...
Backtracking based integer factorisation, primality testing and square root c...Backtracking based integer factorisation, primality testing and square root c...
Backtracking based integer factorisation, primality testing and square root c...csandit
 
Data structures and Big O notation
Data structures and Big O notationData structures and Big O notation
Data structures and Big O notationMuthiah Abbhirami
 
Unit 3-with-privious-quesstions
Unit 3-with-privious-quesstionsUnit 3-with-privious-quesstions
Unit 3-with-privious-quesstionsprabhu teja
 
Lti analysis signal system presentation by vijay ,vishal ,rahul upadhyay ,sat...
Lti analysis signal system presentation by vijay ,vishal ,rahul upadhyay ,sat...Lti analysis signal system presentation by vijay ,vishal ,rahul upadhyay ,sat...
Lti analysis signal system presentation by vijay ,vishal ,rahul upadhyay ,sat...Vijay Munda
 
Recurrence relationships
Recurrence relationshipsRecurrence relationships
Recurrence relationshipsDevansh16
 
Synthesis of linear quantum stochastic systems via quantum feedback networks
Synthesis of linear quantum stochastic systems via quantum feedback networksSynthesis of linear quantum stochastic systems via quantum feedback networks
Synthesis of linear quantum stochastic systems via quantum feedback networkshendrai
 
K neareast neighbor algorithm presentation
K neareast neighbor algorithm presentationK neareast neighbor algorithm presentation
K neareast neighbor algorithm presentationShiraz316
 

What's hot (20)

Mca 4040 analysis and design of algorithm
Mca 4040  analysis and design of algorithmMca 4040  analysis and design of algorithm
Mca 4040 analysis and design of algorithm
 
CS 6402 – DESIGN AND ANALYSIS OF ALGORITHMS questio
CS 6402 – DESIGN AND ANALYSIS OF ALGORITHMS questioCS 6402 – DESIGN AND ANALYSIS OF ALGORITHMS questio
CS 6402 – DESIGN AND ANALYSIS OF ALGORITHMS questio
 
Efficient Sparse Coding Algorithms
Efficient Sparse Coding AlgorithmsEfficient Sparse Coding Algorithms
Efficient Sparse Coding Algorithms
 
ICDE-2015 Shortest Path Traversal Optimization and Analysis for Large Graph C...
ICDE-2015 Shortest Path Traversal Optimization and Analysis for Large Graph C...ICDE-2015 Shortest Path Traversal Optimization and Analysis for Large Graph C...
ICDE-2015 Shortest Path Traversal Optimization and Analysis for Large Graph C...
 
Compositional Program Analysis using Max-SMT
Compositional Program Analysis using Max-SMTCompositional Program Analysis using Max-SMT
Compositional Program Analysis using Max-SMT
 
Computability and Complexity
Computability and ComplexityComputability and Complexity
Computability and Complexity
 
Divide and conquer strategy
Divide and conquer strategyDivide and conquer strategy
Divide and conquer strategy
 
Dynamic programmng2
Dynamic programmng2Dynamic programmng2
Dynamic programmng2
 
How Unstable is an Unstable System
How Unstable is an Unstable SystemHow Unstable is an Unstable System
How Unstable is an Unstable System
 
Complexity
ComplexityComplexity
Complexity
 
Lower bound
Lower boundLower bound
Lower bound
 
Algorithms
AlgorithmsAlgorithms
Algorithms
 
Backtracking based integer factorisation, primality testing and square root c...
Backtracking based integer factorisation, primality testing and square root c...Backtracking based integer factorisation, primality testing and square root c...
Backtracking based integer factorisation, primality testing and square root c...
 
Data structures and Big O notation
Data structures and Big O notationData structures and Big O notation
Data structures and Big O notation
 
Unit 3-with-privious-quesstions
Unit 3-with-privious-quesstionsUnit 3-with-privious-quesstions
Unit 3-with-privious-quesstions
 
Lti analysis signal system presentation by vijay ,vishal ,rahul upadhyay ,sat...
Lti analysis signal system presentation by vijay ,vishal ,rahul upadhyay ,sat...Lti analysis signal system presentation by vijay ,vishal ,rahul upadhyay ,sat...
Lti analysis signal system presentation by vijay ,vishal ,rahul upadhyay ,sat...
 
Lab no.07
Lab no.07Lab no.07
Lab no.07
 
Recurrence relationships
Recurrence relationshipsRecurrence relationships
Recurrence relationships
 
Synthesis of linear quantum stochastic systems via quantum feedback networks
Synthesis of linear quantum stochastic systems via quantum feedback networksSynthesis of linear quantum stochastic systems via quantum feedback networks
Synthesis of linear quantum stochastic systems via quantum feedback networks
 
K neareast neighbor algorithm presentation
K neareast neighbor algorithm presentationK neareast neighbor algorithm presentation
K neareast neighbor algorithm presentation
 

Viewers also liked

OMNI-Prop: Seamless Node Classification on Arbitrary Label Correlation
OMNI-Prop: Seamless Node Classification on Arbitrary Label CorrelationOMNI-Prop: Seamless Node Classification on Arbitrary Label Correlation
OMNI-Prop: Seamless Node Classification on Arbitrary Label CorrelationYuto Yamaguchi
 
Patterns in Interactive Tagging Networks
Patterns in Interactive Tagging NetworksPatterns in Interactive Tagging Networks
Patterns in Interactive Tagging NetworksYuto Yamaguchi
 
Label propagation - Semisupervised Learning with Applications to NLP
Label propagation - Semisupervised Learning with Applications to NLPLabel propagation - Semisupervised Learning with Applications to NLP
Label propagation - Semisupervised Learning with Applications to NLPDavid Przybilla
 
Robust Large-Scale Machine Learning in the Cloud
Robust Large-Scale Machine Learning in the CloudRobust Large-Scale Machine Learning in the Cloud
Robust Large-Scale Machine Learning in the CloudYuto Yamaguchi
 
4 avrachenkov
4 avrachenkov4 avrachenkov
4 avrachenkovYandex
 
CVPR2010: Semi-supervised Learning in Vision: Part 2: Theory
CVPR2010: Semi-supervised Learning in Vision: Part 2: TheoryCVPR2010: Semi-supervised Learning in Vision: Part 2: Theory
CVPR2010: Semi-supervised Learning in Vision: Part 2: Theoryzukun
 
2014.02.13 (Strata) Graph Analysis with One Trillion Edges on Apache Giraph
2014.02.13 (Strata) Graph Analysis with One Trillion Edges on Apache Giraph2014.02.13 (Strata) Graph Analysis with One Trillion Edges on Apache Giraph
2014.02.13 (Strata) Graph Analysis with One Trillion Edges on Apache GiraphAvery Ching
 
Graph based Semi Supervised Learning V1
Graph based Semi Supervised Learning V1Graph based Semi Supervised Learning V1
Graph based Semi Supervised Learning V1Neeta Pande
 
02 probabilistic inference in graphical models
02 probabilistic inference in graphical models02 probabilistic inference in graphical models
02 probabilistic inference in graphical modelszukun
 
Online User Location Inference Exploiting Spatiotemporal Correlations in Soci...
Online User Location Inference Exploiting Spatiotemporal Correlations in Soci...Online User Location Inference Exploiting Spatiotemporal Correlations in Soci...
Online User Location Inference Exploiting Spatiotemporal Correlations in Soci...Yuto Yamaguchi
 
Cyber security and attack analysis : how Cisco uses graph analytics
Cyber security and attack analysis : how Cisco uses graph analyticsCyber security and attack analysis : how Cisco uses graph analytics
Cyber security and attack analysis : how Cisco uses graph analyticsLinkurious
 
Semi supervised learning
Semi supervised learningSemi supervised learning
Semi supervised learningAhmed Taha
 
Community detection in graphs
Community detection in graphsCommunity detection in graphs
Community detection in graphsNicola Barbieri
 
Community Detection in Social Media
Community Detection in Social MediaCommunity Detection in Social Media
Community Detection in Social MediaSymeon Papadopoulos
 
Semi-Supervised Learning
Semi-Supervised LearningSemi-Supervised Learning
Semi-Supervised LearningLukas Tencer
 
Artificial Intelligence 06.3 Bayesian Networks - Belief Propagation - Junctio...
Artificial Intelligence 06.3 Bayesian Networks - Belief Propagation - Junctio...Artificial Intelligence 06.3 Bayesian Networks - Belief Propagation - Junctio...
Artificial Intelligence 06.3 Bayesian Networks - Belief Propagation - Junctio...Andres Mendez-Vazquez
 
Language of Politics on Twitter - 03 Analysis
Language of Politics on Twitter - 03 AnalysisLanguage of Politics on Twitter - 03 Analysis
Language of Politics on Twitter - 03 AnalysisYelena Mejova
 

Viewers also liked (17)

OMNI-Prop: Seamless Node Classification on Arbitrary Label Correlation
OMNI-Prop: Seamless Node Classification on Arbitrary Label CorrelationOMNI-Prop: Seamless Node Classification on Arbitrary Label Correlation
OMNI-Prop: Seamless Node Classification on Arbitrary Label Correlation
 
Patterns in Interactive Tagging Networks
Patterns in Interactive Tagging NetworksPatterns in Interactive Tagging Networks
Patterns in Interactive Tagging Networks
 
Label propagation - Semisupervised Learning with Applications to NLP
Label propagation - Semisupervised Learning with Applications to NLPLabel propagation - Semisupervised Learning with Applications to NLP
Label propagation - Semisupervised Learning with Applications to NLP
 
Robust Large-Scale Machine Learning in the Cloud
Robust Large-Scale Machine Learning in the CloudRobust Large-Scale Machine Learning in the Cloud
Robust Large-Scale Machine Learning in the Cloud
 
4 avrachenkov
4 avrachenkov4 avrachenkov
4 avrachenkov
 
CVPR2010: Semi-supervised Learning in Vision: Part 2: Theory
CVPR2010: Semi-supervised Learning in Vision: Part 2: TheoryCVPR2010: Semi-supervised Learning in Vision: Part 2: Theory
CVPR2010: Semi-supervised Learning in Vision: Part 2: Theory
 
2014.02.13 (Strata) Graph Analysis with One Trillion Edges on Apache Giraph
2014.02.13 (Strata) Graph Analysis with One Trillion Edges on Apache Giraph2014.02.13 (Strata) Graph Analysis with One Trillion Edges on Apache Giraph
2014.02.13 (Strata) Graph Analysis with One Trillion Edges on Apache Giraph
 
Graph based Semi Supervised Learning V1
Graph based Semi Supervised Learning V1Graph based Semi Supervised Learning V1
Graph based Semi Supervised Learning V1
 
02 probabilistic inference in graphical models
02 probabilistic inference in graphical models02 probabilistic inference in graphical models
02 probabilistic inference in graphical models
 
Online User Location Inference Exploiting Spatiotemporal Correlations in Soci...
Online User Location Inference Exploiting Spatiotemporal Correlations in Soci...Online User Location Inference Exploiting Spatiotemporal Correlations in Soci...
Online User Location Inference Exploiting Spatiotemporal Correlations in Soci...
 
Cyber security and attack analysis : how Cisco uses graph analytics
Cyber security and attack analysis : how Cisco uses graph analyticsCyber security and attack analysis : how Cisco uses graph analytics
Cyber security and attack analysis : how Cisco uses graph analytics
 
Semi supervised learning
Semi supervised learningSemi supervised learning
Semi supervised learning
 
Community detection in graphs
Community detection in graphsCommunity detection in graphs
Community detection in graphs
 
Community Detection in Social Media
Community Detection in Social MediaCommunity Detection in Social Media
Community Detection in Social Media
 
Semi-Supervised Learning
Semi-Supervised LearningSemi-Supervised Learning
Semi-Supervised Learning
 
Artificial Intelligence 06.3 Bayesian Networks - Belief Propagation - Junctio...
Artificial Intelligence 06.3 Bayesian Networks - Belief Propagation - Junctio...Artificial Intelligence 06.3 Bayesian Networks - Belief Propagation - Junctio...
Artificial Intelligence 06.3 Bayesian Networks - Belief Propagation - Junctio...
 
Language of Politics on Twitter - 03 Analysis
Language of Politics on Twitter - 03 AnalysisLanguage of Politics on Twitter - 03 Analysis
Language of Politics on Twitter - 03 Analysis
 

Similar to SocNL: Bayesian Label Propagation with Confidence

Introduction to Bayesian Analysis in Python
Introduction to Bayesian Analysis in PythonIntroduction to Bayesian Analysis in Python
Introduction to Bayesian Analysis in PythonPeadar Coyle
 
teteuueieoeofhfhfjffkkkfkfflflflhshssnnvmvvmvv,v,v,nnxmxxm
teteuueieoeofhfhfjffkkkfkfflflflhshssnnvmvvmvv,v,v,nnxmxxmteteuueieoeofhfhfjffkkkfkfflflflhshssnnvmvvmvv,v,v,nnxmxxm
teteuueieoeofhfhfjffkkkfkfflflflhshssnnvmvvmvv,v,v,nnxmxxmzoobiarana76
 
Two strategies for large-scale multi-label classification on the YouTube-8M d...
Two strategies for large-scale multi-label classification on the YouTube-8M d...Two strategies for large-scale multi-label classification on the YouTube-8M d...
Two strategies for large-scale multi-label classification on the YouTube-8M d...Dalei Li
 
DMTM Lecture 03 Regression
DMTM Lecture 03 RegressionDMTM Lecture 03 Regression
DMTM Lecture 03 RegressionPier Luca Lanzi
 
Semi-Supervised Learning with Variational Bayesian Inference and Maximum Unce...
Semi-Supervised Learning with Variational Bayesian Inference and Maximum Unce...Semi-Supervised Learning with Variational Bayesian Inference and Maximum Unce...
Semi-Supervised Learning with Variational Bayesian Inference and Maximum Unce...Kien Duc Do
 
Variational Inference in Python
Variational Inference in PythonVariational Inference in Python
Variational Inference in PythonPeadar Coyle
 
Use of GAN's to analyze chemical reactions
Use of GAN's to analyze chemical reactionsUse of GAN's to analyze chemical reactions
Use of GAN's to analyze chemical reactionsMatthew Clark
 
A new development in the hierarchical clustering of repertory grid data
A new development in the hierarchical clustering of repertory grid dataA new development in the hierarchical clustering of repertory grid data
A new development in the hierarchical clustering of repertory grid dataMark Heckmann
 
ngboost.pptx
ngboost.pptxngboost.pptx
ngboost.pptxHadrian7
 
Flavours of Physics Challenge: Transfer Learning approach
Flavours of Physics Challenge: Transfer Learning approachFlavours of Physics Challenge: Transfer Learning approach
Flavours of Physics Challenge: Transfer Learning approachAlexander Rakhlin
 
Large-Scale Lasso and Elastic-Net Regularized Generalized Linear Models (DB T...
Large-Scale Lasso and Elastic-Net Regularized Generalized Linear Models (DB T...Large-Scale Lasso and Elastic-Net Regularized Generalized Linear Models (DB T...
Large-Scale Lasso and Elastic-Net Regularized Generalized Linear Models (DB T...Spark Summit
 
2015-06-15 Large-Scale Elastic-Net Regularized Generalized Linear Models at S...
2015-06-15 Large-Scale Elastic-Net Regularized Generalized Linear Models at S...2015-06-15 Large-Scale Elastic-Net Regularized Generalized Linear Models at S...
2015-06-15 Large-Scale Elastic-Net Regularized Generalized Linear Models at S...DB Tsai
 
Training Deep Networks with Backprop (D1L4 Insight@DCU Machine Learning Works...
Training Deep Networks with Backprop (D1L4 Insight@DCU Machine Learning Works...Training Deep Networks with Backprop (D1L4 Insight@DCU Machine Learning Works...
Training Deep Networks with Backprop (D1L4 Insight@DCU Machine Learning Works...Universitat Politècnica de Catalunya
 
DMTM Lecture 04 Classification
DMTM Lecture 04 ClassificationDMTM Lecture 04 Classification
DMTM Lecture 04 ClassificationPier Luca Lanzi
 

Similar to SocNL: Bayesian Label Propagation with Confidence (20)

Introduction to Bayesian Analysis in Python
Introduction to Bayesian Analysis in PythonIntroduction to Bayesian Analysis in Python
Introduction to Bayesian Analysis in Python
 
Into to prob_prog_hari
Into to prob_prog_hariInto to prob_prog_hari
Into to prob_prog_hari
 
teteuueieoeofhfhfjffkkkfkfflflflhshssnnvmvvmvv,v,v,nnxmxxm
teteuueieoeofhfhfjffkkkfkfflflflhshssnnvmvvmvv,v,v,nnxmxxmteteuueieoeofhfhfjffkkkfkfflflflhshssnnvmvvmvv,v,v,nnxmxxm
teteuueieoeofhfhfjffkkkfkfflflflhshssnnvmvvmvv,v,v,nnxmxxm
 
Complexity theory
Complexity theory Complexity theory
Complexity theory
 
Two strategies for large-scale multi-label classification on the YouTube-8M d...
Two strategies for large-scale multi-label classification on the YouTube-8M d...Two strategies for large-scale multi-label classification on the YouTube-8M d...
Two strategies for large-scale multi-label classification on the YouTube-8M d...
 
DMTM Lecture 03 Regression
DMTM Lecture 03 RegressionDMTM Lecture 03 Regression
DMTM Lecture 03 Regression
 
Semi-Supervised Learning with Variational Bayesian Inference and Maximum Unce...
Semi-Supervised Learning with Variational Bayesian Inference and Maximum Unce...Semi-Supervised Learning with Variational Bayesian Inference and Maximum Unce...
Semi-Supervised Learning with Variational Bayesian Inference and Maximum Unce...
 
Variational Inference in Python
Variational Inference in PythonVariational Inference in Python
Variational Inference in Python
 
Unit iii update
Unit iii updateUnit iii update
Unit iii update
 
Use of GAN's to analyze chemical reactions
Use of GAN's to analyze chemical reactionsUse of GAN's to analyze chemical reactions
Use of GAN's to analyze chemical reactions
 
Convergence Analysis
Convergence AnalysisConvergence Analysis
Convergence Analysis
 
A new development in the hierarchical clustering of repertory grid data
A new development in the hierarchical clustering of repertory grid dataA new development in the hierarchical clustering of repertory grid data
A new development in the hierarchical clustering of repertory grid data
 
ngboost.pptx
ngboost.pptxngboost.pptx
ngboost.pptx
 
Flavours of Physics Challenge: Transfer Learning approach
Flavours of Physics Challenge: Transfer Learning approachFlavours of Physics Challenge: Transfer Learning approach
Flavours of Physics Challenge: Transfer Learning approach
 
Large-Scale Lasso and Elastic-Net Regularized Generalized Linear Models (DB T...
Large-Scale Lasso and Elastic-Net Regularized Generalized Linear Models (DB T...Large-Scale Lasso and Elastic-Net Regularized Generalized Linear Models (DB T...
Large-Scale Lasso and Elastic-Net Regularized Generalized Linear Models (DB T...
 
2015-06-15 Large-Scale Elastic-Net Regularized Generalized Linear Models at S...
2015-06-15 Large-Scale Elastic-Net Regularized Generalized Linear Models at S...2015-06-15 Large-Scale Elastic-Net Regularized Generalized Linear Models at S...
2015-06-15 Large-Scale Elastic-Net Regularized Generalized Linear Models at S...
 
5954987.ppt
5954987.ppt5954987.ppt
5954987.ppt
 
m7-logic.ppt
m7-logic.pptm7-logic.ppt
m7-logic.ppt
 
Training Deep Networks with Backprop (D1L4 Insight@DCU Machine Learning Works...
Training Deep Networks with Backprop (D1L4 Insight@DCU Machine Learning Works...Training Deep Networks with Backprop (D1L4 Insight@DCU Machine Learning Works...
Training Deep Networks with Backprop (D1L4 Insight@DCU Machine Learning Works...
 
DMTM Lecture 04 Classification
DMTM Lecture 04 ClassificationDMTM Lecture 04 Classification
DMTM Lecture 04 Classification
 

More from Yuto Yamaguchi

When Does Label Propagation Fail? A View from a Network Generative Model@ERAT...
When Does Label Propagation Fail? A View from a Network Generative Model@ERAT...When Does Label Propagation Fail? A View from a Network Generative Model@ERAT...
When Does Label Propagation Fail? A View from a Network Generative Model@ERAT...Yuto Yamaguchi
 
Bridging Relational Learning Algorithms@ビッグデータ基盤勉強会
Bridging Relational Learning Algorithms@ビッグデータ基盤勉強会Bridging Relational Learning Algorithms@ビッグデータ基盤勉強会
Bridging Relational Learning Algorithms@ビッグデータ基盤勉強会Yuto Yamaguchi
 
Tensor Decomposition with Missing Indices
Tensor Decomposition with Missing IndicesTensor Decomposition with Missing Indices
Tensor Decomposition with Missing IndicesYuto Yamaguchi
 
When Does Label Propagation Fail? A View from a Network Generative Model
When Does Label Propagation Fail? A View from a Network Generative ModelWhen Does Label Propagation Fail? A View from a Network Generative Model
When Does Label Propagation Fail? A View from a Network Generative ModelYuto Yamaguchi
 
SIGMOD2013勉強会:Social Media
SIGMOD2013勉強会:Social MediaSIGMOD2013勉強会:Social Media
SIGMOD2013勉強会:Social MediaYuto Yamaguchi
 
Towards Social User Profiling: Unified and Discriminative Influence Model for...
Towards Social User Profiling: Unified and Discriminative Influence Model for...Towards Social User Profiling: Unified and Discriminative Influence Model for...
Towards Social User Profiling: Unified and Discriminative Influence Model for...Yuto Yamaguchi
 
The Length of Bridge Ties: Structural and Geographic Properties of Online So...
The Length of Bridge Ties: Structural and Geographic Properties of Online So...The Length of Bridge Ties: Structural and Geographic Properties of Online So...
The Length of Bridge Ties: Structural and Geographic Properties of Online So...Yuto Yamaguchi
 
WWW2012勉強会:Information Diffusion in Social Networks
WWW2012勉強会:Information Diffusion in Social NetworksWWW2012勉強会:Information Diffusion in Social Networks
WWW2012勉強会:Information Diffusion in Social NetworksYuto Yamaguchi
 
ICDE2012勉強会:Social Media
ICDE2012勉強会:Social MediaICDE2012勉強会:Social Media
ICDE2012勉強会:Social MediaYuto Yamaguchi
 

More from Yuto Yamaguchi (9)

When Does Label Propagation Fail? A View from a Network Generative Model@ERAT...
When Does Label Propagation Fail? A View from a Network Generative Model@ERAT...When Does Label Propagation Fail? A View from a Network Generative Model@ERAT...
When Does Label Propagation Fail? A View from a Network Generative Model@ERAT...
 
Bridging Relational Learning Algorithms@ビッグデータ基盤勉強会
Bridging Relational Learning Algorithms@ビッグデータ基盤勉強会Bridging Relational Learning Algorithms@ビッグデータ基盤勉強会
Bridging Relational Learning Algorithms@ビッグデータ基盤勉強会
 
Tensor Decomposition with Missing Indices
Tensor Decomposition with Missing IndicesTensor Decomposition with Missing Indices
Tensor Decomposition with Missing Indices
 
When Does Label Propagation Fail? A View from a Network Generative Model
When Does Label Propagation Fail? A View from a Network Generative ModelWhen Does Label Propagation Fail? A View from a Network Generative Model
When Does Label Propagation Fail? A View from a Network Generative Model
 
SIGMOD2013勉強会:Social Media
SIGMOD2013勉強会:Social MediaSIGMOD2013勉強会:Social Media
SIGMOD2013勉強会:Social Media
 
Towards Social User Profiling: Unified and Discriminative Influence Model for...
Towards Social User Profiling: Unified and Discriminative Influence Model for...Towards Social User Profiling: Unified and Discriminative Influence Model for...
Towards Social User Profiling: Unified and Discriminative Influence Model for...
 
The Length of Bridge Ties: Structural and Geographic Properties of Online So...
The Length of Bridge Ties: Structural and Geographic Properties of Online So...The Length of Bridge Ties: Structural and Geographic Properties of Online So...
The Length of Bridge Ties: Structural and Geographic Properties of Online So...
 
WWW2012勉強会:Information Diffusion in Social Networks
WWW2012勉強会:Information Diffusion in Social NetworksWWW2012勉強会:Information Diffusion in Social Networks
WWW2012勉強会:Information Diffusion in Social Networks
 
ICDE2012勉強会:Social Media
ICDE2012勉強会:Social MediaICDE2012勉強会:Social Media
ICDE2012勉強会:Social Media
 

Recently uploaded

Artificial intelligence in the post-deep learning era
Artificial intelligence in the post-deep learning eraArtificial intelligence in the post-deep learning era
Artificial intelligence in the post-deep learning eraDeakin University
 
Gen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfGen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfAddepto
 
Key Features Of Token Development (1).pptx
Key  Features Of Token  Development (1).pptxKey  Features Of Token  Development (1).pptx
Key Features Of Token Development (1).pptxLBM Solutions
 
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...shyamraj55
 
Bun (KitWorks Team Study 노별마루 발표 2024.4.22)
Bun (KitWorks Team Study 노별마루 발표 2024.4.22)Bun (KitWorks Team Study 노별마루 발표 2024.4.22)
Bun (KitWorks Team Study 노별마루 발표 2024.4.22)Wonjun Hwang
 
My Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationMy Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationRidwan Fadjar
 
Unleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding ClubUnleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding ClubKalema Edgar
 
Dev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio WebDev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio WebUiPathCommunity
 
Build your next Gen AI Breakthrough - April 2024
Build your next Gen AI Breakthrough - April 2024Build your next Gen AI Breakthrough - April 2024
Build your next Gen AI Breakthrough - April 2024Neo4j
 
Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Enterprise Knowledge
 
Science&tech:THE INFORMATION AGE STS.pdf
Science&tech:THE INFORMATION AGE STS.pdfScience&tech:THE INFORMATION AGE STS.pdf
Science&tech:THE INFORMATION AGE STS.pdfjimielynbastida
 
Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365
Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365
Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 3652toLead Limited
 
Streamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project SetupStreamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project SetupFlorian Wilhelm
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsMark Billinghurst
 
Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?Mattias Andersson
 
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr LapshynFwdays
 
Install Stable Diffusion in windows machine
Install Stable Diffusion in windows machineInstall Stable Diffusion in windows machine
Install Stable Diffusion in windows machinePadma Pradeep
 
Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Commit University
 

Recently uploaded (20)

Artificial intelligence in the post-deep learning era
Artificial intelligence in the post-deep learning eraArtificial intelligence in the post-deep learning era
Artificial intelligence in the post-deep learning era
 
Gen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfGen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdf
 
Vulnerability_Management_GRC_by Sohang Sengupta.pptx
Vulnerability_Management_GRC_by Sohang Sengupta.pptxVulnerability_Management_GRC_by Sohang Sengupta.pptx
Vulnerability_Management_GRC_by Sohang Sengupta.pptx
 
Key Features Of Token Development (1).pptx
Key  Features Of Token  Development (1).pptxKey  Features Of Token  Development (1).pptx
Key Features Of Token Development (1).pptx
 
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
 
DMCC Future of Trade Web3 - Special Edition
DMCC Future of Trade Web3 - Special EditionDMCC Future of Trade Web3 - Special Edition
DMCC Future of Trade Web3 - Special Edition
 
Bun (KitWorks Team Study 노별마루 발표 2024.4.22)
Bun (KitWorks Team Study 노별마루 발표 2024.4.22)Bun (KitWorks Team Study 노별마루 발표 2024.4.22)
Bun (KitWorks Team Study 노별마루 발표 2024.4.22)
 
My Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationMy Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 Presentation
 
Unleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding ClubUnleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding Club
 
Dev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio WebDev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio Web
 
Build your next Gen AI Breakthrough - April 2024
Build your next Gen AI Breakthrough - April 2024Build your next Gen AI Breakthrough - April 2024
Build your next Gen AI Breakthrough - April 2024
 
Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024
 
Science&tech:THE INFORMATION AGE STS.pdf
Science&tech:THE INFORMATION AGE STS.pdfScience&tech:THE INFORMATION AGE STS.pdf
Science&tech:THE INFORMATION AGE STS.pdf
 
Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365
Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365
Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365
 
Streamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project SetupStreamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project Setup
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR Systems
 
Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?
 
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
 
Install Stable Diffusion in windows machine
Install Stable Diffusion in windows machineInstall Stable Diffusion in windows machine
Install Stable Diffusion in windows machine
 
Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!
 

SocNL: Bayesian Label Propagation with Confidence

  • 1. SocNL:  Bayesian  Label   Propaga5on  with  Confidence   Yuto  Yamaguchi†   Christos  Faloutsos‡   Hiroyuki  Kitagawa†     †U.  of  Tsukuba        ‡CMU  
  • 2. Node  Classifica5on 15/01/29 Yuto  Yamaguchi  -­‐  AAAI2015 2 Find: correct labels of unlabeled nodes ? ?
  • 3. Our  focus  –  Classifica5on  confidence   Example  input  graph   Our  intui5on    -­‐  A  is  most  probably  conserva5ve    -­‐  B  may  be  conserva5ve   person   è It’s  good  to  have  confidence  for  our  predic5on        e.g.,  A  is  conserva5ve  with  confidence  score  0.9            B  is  conserva5ve  with  confidence  score  0.55  
  • 4. Contribu5ons   •  Novel  Algorithm   –  Simple,  fast,  and  incorporates  confidence   •  Theore5cal  Analysis   –  Convergence  guarantee  &  speed   –  Equivalence  to  LP  and  Bayesian  inference   •  Empirical  Analysis   –  Higher  accuracy  than  compe5tors   –  Three  different  real  network  datasets  
  • 6. Smoothness  assump5on   (widely  adopted)    Connected  nodes  are  likely  to  share  a  label   B A
  • 7. Our  Idea   B A Smoothness  assump5on  +  confidence    IF  a  node  has  a  lot  of  red/blue  neighbors    THEN  we  can  confidently  say  that  it  is  red/blue   Confident   Not  confident   More  evidence,  more  confidence    à    Bayesian  inference  
  • 8. Cases  to  consider   •  Case1:  without  unlabeled  neighbors   – Easy  but  unrealis5c   •  Case2:  with  unlabeled  neighbors   – We  want  to  handle  this  case   ? ?   ?  
  • 9. Case1:  No  unlabeled  neighbors   ? Prior   knowledge   evidence   +   Result   Detail   DCM  (Dirichlet  compound  mul5nomial)  leads  to  simple  result:   ∝   fik:  probability  that  i  has  label  k   nik:  number  of  i’s  neighbors  with  label  k   αk:  prior    
  • 10. Case2:  With  unlabeled  neighbors   A   B   Classifica>on  result  for  A  affects  B       Classifica>on  result  for  B  affects  A   In  this  case  we  need  to  solve  the  recursive  equa5on:   Aij:  entry  of  adjacency  matrix   Detail  
  • 11. Yes,  we  can  solve  it   (Please  see  the  paper  for  detail)   •  Simple:  We  just  need  to  do  matrix  inversion   •  Fast:  power  itera5on  for  sparse  matrix  inversion                        is  fast  (PUU  is  sparse)   •  Confidence:  this  equa5on  is  from  Bayesian  inference  
  • 13. Convergence  guarantee  &  speed   Theorem  1:   The  itera5ve  algorithm  of  SocNL  always  converges  on  arbitrary   graphs  if  use  non-­‐zero  prior  values   Theorem  2:   SocNL  converges  faster  if  use  larger  prior  values     Theorem  3:   Time  complexity  of  each  itera5on  of  SocNL  is  O(  K(N+M)  )  
  • 14. Equivalence   Theorem  4:   SocNL  is  equivalent  to  normal  LP   if  uses  prior  values  =  0   Theorem  5:   SocNL  is  equivalent  to  Bayesian  inference  over  DCM   if  ignores  unlabeled  nodes   *  DCM:  Dirichlet  compound  mul5nomial  
  • 16. Experimental  seings   ○  Datasets   ○  Compe5tors   •  Label  Propaga>on  [ICML03]   •  Myopic:  SocNL  ignoring  unlabeled  nodes  
  • 17. Results   Myopic  not  good  L   SocNL  shows  higher  overall  accuracy   than  compe>tors    J   Upper  is  be3er     Myopic  not  good  L   LP  not  good  L  
  • 18. Summary   •  Proposed  SocNL   –  Simple,  fast,  and  incorporates  confidence   •  Theore5cal  Analysis   –  Convergence  (Theorems  1,2,3)   –  Equivalence  (Theorems  4,5)   •  Empirical  Analysis   –  Higher  overall  accuracy  than  compe5tors   Upper  is  be3er