SlideShare a Scribd company logo
1 of 21
Download to read offline
A Unifying Probabilistic Perspective for Spectral
Dimensionality Reduction: Insights and New
Models – Neil D. Lawrence
Presenters: Sean Golliher and Derek Reimanis
1 / 21
Claims and Contributions
“Unifying Approach”: Views previous methods as less general that
this approach and proves they are special cases
Presents the Unified Algorithm: Maximum Entropy Unfolding (MEU)
Improves Local Linear Embedding (LLE) and introduces ALLE
(Acyclic Version)
Introduces 3rd Algorithm DRILL. Estimating the dimensional
structure from data (rather than K nearest neighbors)
Mostly theoretical paper
2 / 21
Features of Spectral Methods
Spectral Methods: Low dimensional representations are derived from
eigenvectors of specially constructed matrices.
Methods Share a Common Approach:
Compute nearest neighbors of input pattern
Construct a weighted graph based on neighborhood relations
Derive a matrix from the weighted graph
Produce a representation of the matrix from the eigenvectors of the
matrix
Keep this mind when trying to understand unification
3 / 21
Most Common Examples
PCA: Technique that preserves the covariance of the data structure
MDS Multidimensional Scaling: preserves inner products of input
patterns. ex: xi − xj 2
ISOMAP: Preserves pairwise distances by measuring along a
sub-manifold from the sample space. Variant of MDS which uses
geodesics along sub-manifold
geodesic: curve whose tangent vectors remain parallel if they are
transported along it
4 / 21
Local Linear Embedding (LLE)
Tries to capture non-linear relationships by preserving local linear
structure of data.
Four Step Algorithm:
Step 1: Compute K-nearest neighbors of each higher-dimensional input
pattern xi and create directed graph whose edges indicate nearest
neighbor relations.
Step 2: Assign weights Wj to edges in graph. Each input pattern, and
neighbors, can be viewed as samples from linear patch on a lower
dimensional sub manifold.
Find the “reconstruction weights” that provide a local linear fit of the
k + 1 one points at each of the n points in the data set. Minimize
error function:
E(W|D) =
x∈D
x −
y∈N(x)
Wxyy 2
5 / 21
Local Linear Embedding (LLE) Cont’d
Step 3: After n local models are constructed keep weights fixed and
find new projection. We turn this into a minimization problem similar
to MDS
E(Z|D) =
x∈D
zx
−
y∈N(x)
Wxyzy 2
Put into a matrix Mxy that gives us a new objective function in
matrix form (error function):
E(Z|W ) =
x,y
Mxy(zx
)T
zy
Find the m + 1 eigenvectors with the smallest eigenvectors and drop
them. Giving us a new set of coordinates for new k-dimensional space.
6 / 21
Maximum Entropy Unfolding (MEU)
Kernel PCA idea: Apply “Kernal Trick” to transform to higher
dimensional space. Then transform using covariance space as in PCA.
Increases the features space rather than reducing the dimensions.
This motived the development of Maximum Variance Unfolding
(MVU) .
MEU: Since entropy is related to variance they use entropy and
obtain a probabilistic model. Derive a density over p(Y) directly (not
over squared distances) by constraining the expected square
inter-point distances dij of any two samples (yi ) and (yj )
7 / 21
Maximum Entropy Unfolding (MEU) Cont’d
Observations are over squared distances they derive a density that
gives rise to those distances p(Y).
Uses KL divergence for calculating entropy between two distributions.
However, we need two distributions: H = − p(Y)log p(Y)
m(Y)dY
Assume a very broad, spherical, Gaussian density that can be
assumed to be small.
Solving for Y the density that minimizes the KL divergence under the
constraints on the expectations is:
p(Y) =
p
j=1
|L+γI|1/2
τn/2 exp(−1/2yT
j (L + γI))yj
Where L is a special matrix where off diagonal elements contain
information for D. Similar to previous examples.
Key point is that this assumes independence of density across data
features (multiplying densities)
8 / 21
Maximum Entropy Unfolding (MEU) Cont’d
GMRF: (finite-dimensional) random vector following a multivariate
normal (or Gaussian) distribution.
Since GRFs show provide an alternative approach to reducing number
of parameters in covariance matrix they use GRF approach which is
common on spatial methods.
Independence in their model is being expressed over data features
instead of data points. This assumes features are i.i.d
For their model the number of parameters does not increase with the
number of data. As number of features increases their is a clear
“blessing of dimensionality”.
9 / 21
Maximum Entropy Unfolding (MEU) Cont’d
With the GRF representation they show that with some manipulation
they can compute < dij >from the covariance matrix:
< dij >= p/2(ki,j − 2ki,j + ki,j )
This is of the same form as distances and similarities in Kernel PCA
methods
10 / 21
MEU and LLE
They show that LLE is approximating maximum likelihood in their
MEU model
Finding estimates of the parameters that maximize the probability of
observing the data that we have
Thus the claim of a “unifying model”. LLE is special case of MEU.
Also showed that using pseudo-likelihood in the MEU model reduces
to the generalization they presented in equation (9) for LLE.
Using fact that join probability density of GRF can be represented as
a factorization of the cliques of the graph.
Pseudo-likelihood reduces the multiplication of the of the distribution
presented earlier to the objective matrix for LLE.
11 / 21
Acyclic Locally Linear Embedding (ALLE)
Pseudo-likelihood approximation to the joint probability distribution
of a collection of random variables.
Thus the claim of a “unifying model”. LLE is special case of MEU.
Also showed that using pseudo-likelihood in the MEU model reduces
to the generalization they presented in equation (9) for LLE.
If they force their matrix M to be lower triangular they that the
covariance of the try log-likelihood of y log p(Y) factors into n
independent regression functions.
This approach requires an ordering of the data because of the
restriction for j > i in the matrix M
12 / 21
Dimensionality reduction through Regularization of the
Inverse covariance in the Log Likelihood (DRILL)
Optimization technique that applies L-1 regularization of
dependencies to estimate graph structure
How does this compare to the nearest-neighbor approaches?
E(Λ) = − log p(Y ) +
i<j
λi,j 1
Maximize E(Λ) through Least Absolute Shrinkage and Selection
Operator (LASSO) regression
13 / 21
Gaussian process latent variable model (GP-LVM)
likelihood
Process which maps data from a latent space to a non-latent data
space where the locale of points is determined by maximizing a
Gaussian likelihood.
This process allows us to map from a low dimensional space Rd
(latent space) to a high dimensional space RD (data space)
GP-LVM scoring
Define hyperparameter θ = (σ, θrbf , θnoise)
lml = max
σ,θrbf ,θnoise
log p(Y |X, θ)
Find hyperparameter values through gradient descent
14 / 21
Motion Capture Experiments
15 / 21
Motion Capture Experiments
16 / 21
Robot Navigation Experiments
17 / 21
Robot Navigation Experiments
18 / 21
GP-LVM scores for experiments
19 / 21
Structural learning for DRILL
20 / 21
Discussion Questions
Is there an equivalent “no free lunch theorem” for dimensionality
reduction? Why or why not?
Were you convinced that this is a unifying approach?
If this is a unifying approach why is MEU the worst performer?
Does a theoretical paper such as this need more experiments?
Is the GP-LVM score a good performance measure for comparing
these methods?
Are the experiments themselves introducing any bias in the results?
Why do you think ALLE outperforms MEU?
Is the “blessing of dimensionality” a valid claim?
21 / 21

More Related Content

What's hot

A review on structure learning in GNN
A review on structure learning in GNNA review on structure learning in GNN
A review on structure learning in GNNtuxette
 
From RNN to neural networks for cyclic undirected graphs
From RNN to neural networks for cyclic undirected graphsFrom RNN to neural networks for cyclic undirected graphs
From RNN to neural networks for cyclic undirected graphstuxette
 
High Dimensional Data Visualization using t-SNE
High Dimensional Data Visualization using t-SNEHigh Dimensional Data Visualization using t-SNE
High Dimensional Data Visualization using t-SNEKai-Wen Zhao
 
Dimensionality reduction with UMAP
Dimensionality reduction with UMAPDimensionality reduction with UMAP
Dimensionality reduction with UMAPJakub Bartczuk
 
Visualizing Data Using t-SNE
Visualizing Data Using t-SNEVisualizing Data Using t-SNE
Visualizing Data Using t-SNEDavid Khosid
 
How to Layer a Directed Acyclic Graph (GD 2001)
How to Layer a Directed Acyclic Graph (GD 2001)How to Layer a Directed Acyclic Graph (GD 2001)
How to Layer a Directed Acyclic Graph (GD 2001)Nikola S. Nikolov
 
International Journal of Engineering Research and Development
International Journal of Engineering Research and DevelopmentInternational Journal of Engineering Research and Development
International Journal of Engineering Research and DevelopmentIJERD Editor
 
Double Patterning
Double PatterningDouble Patterning
Double PatterningDanny Luk
 
2012 mdsp pr04 monte carlo
2012 mdsp pr04 monte carlo2012 mdsp pr04 monte carlo
2012 mdsp pr04 monte carlonozomuhamada
 
2012 mdsp pr08 nonparametric approach
2012 mdsp pr08 nonparametric approach2012 mdsp pr08 nonparametric approach
2012 mdsp pr08 nonparametric approachnozomuhamada
 
Generalized Notions of Data Depth
Generalized Notions of Data DepthGeneralized Notions of Data Depth
Generalized Notions of Data DepthMukund Raj
 
Satellite image compression technique
Satellite image compression techniqueSatellite image compression technique
Satellite image compression techniqueacijjournal
 
TENSOR VOTING BASED BINARY CLASSIFIER
TENSOR VOTING BASED BINARY CLASSIFIERTENSOR VOTING BASED BINARY CLASSIFIER
TENSOR VOTING BASED BINARY CLASSIFIERcscpconf
 
Spme 2013 segmentation
Spme 2013 segmentationSpme 2013 segmentation
Spme 2013 segmentationQujiang Lei
 

What's hot (20)

A review on structure learning in GNN
A review on structure learning in GNNA review on structure learning in GNN
A review on structure learning in GNN
 
03 Data Representation
03 Data Representation03 Data Representation
03 Data Representation
 
From RNN to neural networks for cyclic undirected graphs
From RNN to neural networks for cyclic undirected graphsFrom RNN to neural networks for cyclic undirected graphs
From RNN to neural networks for cyclic undirected graphs
 
High Dimensional Data Visualization using t-SNE
High Dimensional Data Visualization using t-SNEHigh Dimensional Data Visualization using t-SNE
High Dimensional Data Visualization using t-SNE
 
Dimensionality reduction with UMAP
Dimensionality reduction with UMAPDimensionality reduction with UMAP
Dimensionality reduction with UMAP
 
06 Vector Visualization
06 Vector Visualization06 Vector Visualization
06 Vector Visualization
 
Visualizing Data Using t-SNE
Visualizing Data Using t-SNEVisualizing Data Using t-SNE
Visualizing Data Using t-SNE
 
reportVPLProject
reportVPLProjectreportVPLProject
reportVPLProject
 
05 Scalar Visualization
05 Scalar Visualization05 Scalar Visualization
05 Scalar Visualization
 
regions
regionsregions
regions
 
How to Layer a Directed Acyclic Graph (GD 2001)
How to Layer a Directed Acyclic Graph (GD 2001)How to Layer a Directed Acyclic Graph (GD 2001)
How to Layer a Directed Acyclic Graph (GD 2001)
 
International Journal of Engineering Research and Development
International Journal of Engineering Research and DevelopmentInternational Journal of Engineering Research and Development
International Journal of Engineering Research and Development
 
Double Patterning
Double PatterningDouble Patterning
Double Patterning
 
2012 mdsp pr04 monte carlo
2012 mdsp pr04 monte carlo2012 mdsp pr04 monte carlo
2012 mdsp pr04 monte carlo
 
2012 mdsp pr08 nonparametric approach
2012 mdsp pr08 nonparametric approach2012 mdsp pr08 nonparametric approach
2012 mdsp pr08 nonparametric approach
 
Generalized Notions of Data Depth
Generalized Notions of Data DepthGeneralized Notions of Data Depth
Generalized Notions of Data Depth
 
Em molnar2015
Em molnar2015Em molnar2015
Em molnar2015
 
Satellite image compression technique
Satellite image compression techniqueSatellite image compression technique
Satellite image compression technique
 
TENSOR VOTING BASED BINARY CLASSIFIER
TENSOR VOTING BASED BINARY CLASSIFIERTENSOR VOTING BASED BINARY CLASSIFIER
TENSOR VOTING BASED BINARY CLASSIFIER
 
Spme 2013 segmentation
Spme 2013 segmentationSpme 2013 segmentation
Spme 2013 segmentation
 

Similar to A Unifying Probabilistic Perspective for Spectral Dimensionality Reduction

Reweighting and Boosting to uniforimty in HEP
Reweighting and Boosting to uniforimty in HEPReweighting and Boosting to uniforimty in HEP
Reweighting and Boosting to uniforimty in HEParogozhnikov
 
An_Accelerated_Nearest_Neighbor_Search_Method_for_the_K-Means_Clustering_Algo...
An_Accelerated_Nearest_Neighbor_Search_Method_for_the_K-Means_Clustering_Algo...An_Accelerated_Nearest_Neighbor_Search_Method_for_the_K-Means_Clustering_Algo...
An_Accelerated_Nearest_Neighbor_Search_Method_for_the_K-Means_Clustering_Algo...Adam Fausett
 
Graph Neural Network in practice
Graph Neural Network in practiceGraph Neural Network in practice
Graph Neural Network in practicetuxette
 
A Robust Method Based On LOVO Functions For Solving Least Squares Problems
A Robust Method Based On LOVO Functions For Solving Least Squares ProblemsA Robust Method Based On LOVO Functions For Solving Least Squares Problems
A Robust Method Based On LOVO Functions For Solving Least Squares ProblemsDawn Cook
 
A new implementation of k-MLE for mixture modelling of Wishart distributions
A new implementation of k-MLE for mixture modelling of Wishart distributionsA new implementation of k-MLE for mixture modelling of Wishart distributions
A new implementation of k-MLE for mixture modelling of Wishart distributionsFrank Nielsen
 
Matrix Completion Presentation
Matrix Completion PresentationMatrix Completion Presentation
Matrix Completion PresentationMichael Hankin
 
Investigation on the Pattern Synthesis of Subarray Weights for Low EMI Applic...
Investigation on the Pattern Synthesis of Subarray Weights for Low EMI Applic...Investigation on the Pattern Synthesis of Subarray Weights for Low EMI Applic...
Investigation on the Pattern Synthesis of Subarray Weights for Low EMI Applic...IOSRJECE
 
2014 vulnerability assesment of spatial network - models and solutions
2014   vulnerability assesment of spatial network - models and solutions2014   vulnerability assesment of spatial network - models and solutions
2014 vulnerability assesment of spatial network - models and solutionsFrancisco Pérez
 
Iterative Soft Decision Based Complex K-best MIMO Decoder
Iterative Soft Decision Based Complex K-best MIMO DecoderIterative Soft Decision Based Complex K-best MIMO Decoder
Iterative Soft Decision Based Complex K-best MIMO DecoderCSCJournals
 
Iterative Soft Decision Based Complex K-best MIMO Decoder
Iterative Soft Decision Based Complex K-best MIMO DecoderIterative Soft Decision Based Complex K-best MIMO Decoder
Iterative Soft Decision Based Complex K-best MIMO DecoderCSCJournals
 
Image Processing
Image ProcessingImage Processing
Image ProcessingTuyen Pham
 
Singh gordon-unified-factorization-ecml
Singh gordon-unified-factorization-ecmlSingh gordon-unified-factorization-ecml
Singh gordon-unified-factorization-ecmlHuỳnh Thông
 
COMPARISON OF VOLUME AND DISTANCE CONSTRAINT ON HYPERSPECTRAL UNMIXING
COMPARISON OF VOLUME AND DISTANCE CONSTRAINT ON HYPERSPECTRAL UNMIXINGCOMPARISON OF VOLUME AND DISTANCE CONSTRAINT ON HYPERSPECTRAL UNMIXING
COMPARISON OF VOLUME AND DISTANCE CONSTRAINT ON HYPERSPECTRAL UNMIXINGcsandit
 
Face recognition using laplacianfaces (synopsis)
Face recognition using laplacianfaces (synopsis)Face recognition using laplacianfaces (synopsis)
Face recognition using laplacianfaces (synopsis)Mumbai Academisc
 

Similar to A Unifying Probabilistic Perspective for Spectral Dimensionality Reduction (20)

Reweighting and Boosting to uniforimty in HEP
Reweighting and Boosting to uniforimty in HEPReweighting and Boosting to uniforimty in HEP
Reweighting and Boosting to uniforimty in HEP
 
10.1.1.630.8055
10.1.1.630.805510.1.1.630.8055
10.1.1.630.8055
 
2009 asilomar
2009 asilomar2009 asilomar
2009 asilomar
 
17_monte_carlo.pdf
17_monte_carlo.pdf17_monte_carlo.pdf
17_monte_carlo.pdf
 
An_Accelerated_Nearest_Neighbor_Search_Method_for_the_K-Means_Clustering_Algo...
An_Accelerated_Nearest_Neighbor_Search_Method_for_the_K-Means_Clustering_Algo...An_Accelerated_Nearest_Neighbor_Search_Method_for_the_K-Means_Clustering_Algo...
An_Accelerated_Nearest_Neighbor_Search_Method_for_the_K-Means_Clustering_Algo...
 
Graph Neural Network in practice
Graph Neural Network in practiceGraph Neural Network in practice
Graph Neural Network in practice
 
A Robust Method Based On LOVO Functions For Solving Least Squares Problems
A Robust Method Based On LOVO Functions For Solving Least Squares ProblemsA Robust Method Based On LOVO Functions For Solving Least Squares Problems
A Robust Method Based On LOVO Functions For Solving Least Squares Problems
 
A new implementation of k-MLE for mixture modelling of Wishart distributions
A new implementation of k-MLE for mixture modelling of Wishart distributionsA new implementation of k-MLE for mixture modelling of Wishart distributions
A new implementation of k-MLE for mixture modelling of Wishart distributions
 
Matrix Completion Presentation
Matrix Completion PresentationMatrix Completion Presentation
Matrix Completion Presentation
 
Investigation on the Pattern Synthesis of Subarray Weights for Low EMI Applic...
Investigation on the Pattern Synthesis of Subarray Weights for Low EMI Applic...Investigation on the Pattern Synthesis of Subarray Weights for Low EMI Applic...
Investigation on the Pattern Synthesis of Subarray Weights for Low EMI Applic...
 
QMC: Transition Workshop - Small Sample Statistical Analysis and Algorithms f...
QMC: Transition Workshop - Small Sample Statistical Analysis and Algorithms f...QMC: Transition Workshop - Small Sample Statistical Analysis and Algorithms f...
QMC: Transition Workshop - Small Sample Statistical Analysis and Algorithms f...
 
2014 vulnerability assesment of spatial network - models and solutions
2014   vulnerability assesment of spatial network - models and solutions2014   vulnerability assesment of spatial network - models and solutions
2014 vulnerability assesment of spatial network - models and solutions
 
Iterative Soft Decision Based Complex K-best MIMO Decoder
Iterative Soft Decision Based Complex K-best MIMO DecoderIterative Soft Decision Based Complex K-best MIMO Decoder
Iterative Soft Decision Based Complex K-best MIMO Decoder
 
Iterative Soft Decision Based Complex K-best MIMO Decoder
Iterative Soft Decision Based Complex K-best MIMO DecoderIterative Soft Decision Based Complex K-best MIMO Decoder
Iterative Soft Decision Based Complex K-best MIMO Decoder
 
Line
LineLine
Line
 
(Slide)concentration effect
(Slide)concentration effect(Slide)concentration effect
(Slide)concentration effect
 
Image Processing
Image ProcessingImage Processing
Image Processing
 
Singh gordon-unified-factorization-ecml
Singh gordon-unified-factorization-ecmlSingh gordon-unified-factorization-ecml
Singh gordon-unified-factorization-ecml
 
COMPARISON OF VOLUME AND DISTANCE CONSTRAINT ON HYPERSPECTRAL UNMIXING
COMPARISON OF VOLUME AND DISTANCE CONSTRAINT ON HYPERSPECTRAL UNMIXINGCOMPARISON OF VOLUME AND DISTANCE CONSTRAINT ON HYPERSPECTRAL UNMIXING
COMPARISON OF VOLUME AND DISTANCE CONSTRAINT ON HYPERSPECTRAL UNMIXING
 
Face recognition using laplacianfaces (synopsis)
Face recognition using laplacianfaces (synopsis)Face recognition using laplacianfaces (synopsis)
Face recognition using laplacianfaces (synopsis)
 

More from Sean Golliher

Time Series Forecasting using Neural Nets (GNNNs)
Time Series Forecasting using Neural Nets (GNNNs)Time Series Forecasting using Neural Nets (GNNNs)
Time Series Forecasting using Neural Nets (GNNNs)Sean Golliher
 
Property Matching and Query Expansion on Linked Data Using Kullback-Leibler D...
Property Matching and Query Expansion on Linked Data Using Kullback-Leibler D...Property Matching and Query Expansion on Linked Data Using Kullback-Leibler D...
Property Matching and Query Expansion on Linked Data Using Kullback-Leibler D...Sean Golliher
 
Lecture 9 - Machine Learning and Support Vector Machines (SVM)
Lecture 9 - Machine Learning and Support Vector Machines (SVM)Lecture 9 - Machine Learning and Support Vector Machines (SVM)
Lecture 9 - Machine Learning and Support Vector Machines (SVM)Sean Golliher
 
Probabilistic Retrieval Models - Sean Golliher Lecture 8 MSU CSCI 494
Probabilistic Retrieval Models - Sean Golliher Lecture 8 MSU CSCI 494Probabilistic Retrieval Models - Sean Golliher Lecture 8 MSU CSCI 494
Probabilistic Retrieval Models - Sean Golliher Lecture 8 MSU CSCI 494Sean Golliher
 
Lecture 7- Text Statistics and Document Parsing
Lecture 7- Text Statistics and Document ParsingLecture 7- Text Statistics and Document Parsing
Lecture 7- Text Statistics and Document ParsingSean Golliher
 
Information Retrieval, Encoding, Indexing, Big Table. Lecture 6 - Indexing
Information Retrieval, Encoding, Indexing, Big Table. Lecture 6  - IndexingInformation Retrieval, Encoding, Indexing, Big Table. Lecture 6  - Indexing
Information Retrieval, Encoding, Indexing, Big Table. Lecture 6 - IndexingSean Golliher
 
PageRank and The Google Matrix
PageRank and The Google MatrixPageRank and The Google Matrix
PageRank and The Google MatrixSean Golliher
 
CSCI 494 - Lect. 3. Anatomy of Search Engines/Building a Crawler
CSCI 494 - Lect. 3. Anatomy of Search Engines/Building a CrawlerCSCI 494 - Lect. 3. Anatomy of Search Engines/Building a Crawler
CSCI 494 - Lect. 3. Anatomy of Search Engines/Building a CrawlerSean Golliher
 

More from Sean Golliher (9)

Time Series Forecasting using Neural Nets (GNNNs)
Time Series Forecasting using Neural Nets (GNNNs)Time Series Forecasting using Neural Nets (GNNNs)
Time Series Forecasting using Neural Nets (GNNNs)
 
Goprez sg
Goprez  sgGoprez  sg
Goprez sg
 
Property Matching and Query Expansion on Linked Data Using Kullback-Leibler D...
Property Matching and Query Expansion on Linked Data Using Kullback-Leibler D...Property Matching and Query Expansion on Linked Data Using Kullback-Leibler D...
Property Matching and Query Expansion on Linked Data Using Kullback-Leibler D...
 
Lecture 9 - Machine Learning and Support Vector Machines (SVM)
Lecture 9 - Machine Learning and Support Vector Machines (SVM)Lecture 9 - Machine Learning and Support Vector Machines (SVM)
Lecture 9 - Machine Learning and Support Vector Machines (SVM)
 
Probabilistic Retrieval Models - Sean Golliher Lecture 8 MSU CSCI 494
Probabilistic Retrieval Models - Sean Golliher Lecture 8 MSU CSCI 494Probabilistic Retrieval Models - Sean Golliher Lecture 8 MSU CSCI 494
Probabilistic Retrieval Models - Sean Golliher Lecture 8 MSU CSCI 494
 
Lecture 7- Text Statistics and Document Parsing
Lecture 7- Text Statistics and Document ParsingLecture 7- Text Statistics and Document Parsing
Lecture 7- Text Statistics and Document Parsing
 
Information Retrieval, Encoding, Indexing, Big Table. Lecture 6 - Indexing
Information Retrieval, Encoding, Indexing, Big Table. Lecture 6  - IndexingInformation Retrieval, Encoding, Indexing, Big Table. Lecture 6  - Indexing
Information Retrieval, Encoding, Indexing, Big Table. Lecture 6 - Indexing
 
PageRank and The Google Matrix
PageRank and The Google MatrixPageRank and The Google Matrix
PageRank and The Google Matrix
 
CSCI 494 - Lect. 3. Anatomy of Search Engines/Building a Crawler
CSCI 494 - Lect. 3. Anatomy of Search Engines/Building a CrawlerCSCI 494 - Lect. 3. Anatomy of Search Engines/Building a Crawler
CSCI 494 - Lect. 3. Anatomy of Search Engines/Building a Crawler
 

Recently uploaded

OECD bibliometric indicators: Selected highlights, April 2024
OECD bibliometric indicators: Selected highlights, April 2024OECD bibliometric indicators: Selected highlights, April 2024
OECD bibliometric indicators: Selected highlights, April 2024innovationoecd
 
(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)
(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)
(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)riyaescorts54
 
User Guide: Pulsar™ Weather Station (Columbia Weather Systems)
User Guide: Pulsar™ Weather Station (Columbia Weather Systems)User Guide: Pulsar™ Weather Station (Columbia Weather Systems)
User Guide: Pulsar™ Weather Station (Columbia Weather Systems)Columbia Weather Systems
 
Dubai Calls Girl Lisa O525547819 Lexi Call Girls In Dubai
Dubai Calls Girl Lisa O525547819 Lexi Call Girls In DubaiDubai Calls Girl Lisa O525547819 Lexi Call Girls In Dubai
Dubai Calls Girl Lisa O525547819 Lexi Call Girls In Dubaikojalkojal131
 
STOPPED FLOW METHOD & APPLICATION MURUGAVENI B.pptx
STOPPED FLOW METHOD & APPLICATION MURUGAVENI B.pptxSTOPPED FLOW METHOD & APPLICATION MURUGAVENI B.pptx
STOPPED FLOW METHOD & APPLICATION MURUGAVENI B.pptxMurugaveni B
 
Behavioral Disorder: Schizophrenia & it's Case Study.pdf
Behavioral Disorder: Schizophrenia & it's Case Study.pdfBehavioral Disorder: Schizophrenia & it's Case Study.pdf
Behavioral Disorder: Schizophrenia & it's Case Study.pdfSELF-EXPLANATORY
 
Harmful and Useful Microorganisms Presentation
Harmful and Useful Microorganisms PresentationHarmful and Useful Microorganisms Presentation
Harmful and Useful Microorganisms Presentationtahreemzahra82
 
Four Spheres of the Earth Presentation.ppt
Four Spheres of the Earth Presentation.pptFour Spheres of the Earth Presentation.ppt
Four Spheres of the Earth Presentation.pptJoemSTuliba
 
Call Girls In Nihal Vihar Delhi ❤️8860477959 Looking Escorts In 24/7 Delhi NCR
Call Girls In Nihal Vihar Delhi ❤️8860477959 Looking Escorts In 24/7 Delhi NCRCall Girls In Nihal Vihar Delhi ❤️8860477959 Looking Escorts In 24/7 Delhi NCR
Call Girls In Nihal Vihar Delhi ❤️8860477959 Looking Escorts In 24/7 Delhi NCRlizamodels9
 
Pests of Bengal gram_Identification_Dr.UPR.pdf
Pests of Bengal gram_Identification_Dr.UPR.pdfPests of Bengal gram_Identification_Dr.UPR.pdf
Pests of Bengal gram_Identification_Dr.UPR.pdfPirithiRaju
 
Citronella presentation SlideShare mani upadhyay
Citronella presentation SlideShare mani upadhyayCitronella presentation SlideShare mani upadhyay
Citronella presentation SlideShare mani upadhyayupadhyaymani499
 
Pests of jatropha_Bionomics_identification_Dr.UPR.pdf
Pests of jatropha_Bionomics_identification_Dr.UPR.pdfPests of jatropha_Bionomics_identification_Dr.UPR.pdf
Pests of jatropha_Bionomics_identification_Dr.UPR.pdfPirithiRaju
 
《Queensland毕业文凭-昆士兰大学毕业证成绩单》
《Queensland毕业文凭-昆士兰大学毕业证成绩单》《Queensland毕业文凭-昆士兰大学毕业证成绩单》
《Queensland毕业文凭-昆士兰大学毕业证成绩单》rnrncn29
 
Best Call Girls In Sector 29 Gurgaon❤️8860477959 EscorTs Service In 24/7 Delh...
Best Call Girls In Sector 29 Gurgaon❤️8860477959 EscorTs Service In 24/7 Delh...Best Call Girls In Sector 29 Gurgaon❤️8860477959 EscorTs Service In 24/7 Delh...
Best Call Girls In Sector 29 Gurgaon❤️8860477959 EscorTs Service In 24/7 Delh...lizamodels9
 
preservation, maintanence and improvement of industrial organism.pptx
preservation, maintanence and improvement of industrial organism.pptxpreservation, maintanence and improvement of industrial organism.pptx
preservation, maintanence and improvement of industrial organism.pptxnoordubaliya2003
 
BIOETHICS IN RECOMBINANT DNA TECHNOLOGY.
BIOETHICS IN RECOMBINANT DNA TECHNOLOGY.BIOETHICS IN RECOMBINANT DNA TECHNOLOGY.
BIOETHICS IN RECOMBINANT DNA TECHNOLOGY.PraveenaKalaiselvan1
 
Microphone- characteristics,carbon microphone, dynamic microphone.pptx
Microphone- characteristics,carbon microphone, dynamic microphone.pptxMicrophone- characteristics,carbon microphone, dynamic microphone.pptx
Microphone- characteristics,carbon microphone, dynamic microphone.pptxpriyankatabhane
 
Environmental Biotechnology Topic:- Microbial Biosensor
Environmental Biotechnology Topic:- Microbial BiosensorEnvironmental Biotechnology Topic:- Microbial Biosensor
Environmental Biotechnology Topic:- Microbial Biosensorsonawaneprad
 
Carbon Dioxide Capture and Storage (CSS)
Carbon Dioxide Capture and Storage (CSS)Carbon Dioxide Capture and Storage (CSS)
Carbon Dioxide Capture and Storage (CSS)Tamer Koksalan, PhD
 

Recently uploaded (20)

OECD bibliometric indicators: Selected highlights, April 2024
OECD bibliometric indicators: Selected highlights, April 2024OECD bibliometric indicators: Selected highlights, April 2024
OECD bibliometric indicators: Selected highlights, April 2024
 
(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)
(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)
(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)
 
Hot Sexy call girls in Moti Nagar,🔝 9953056974 🔝 escort Service
Hot Sexy call girls in  Moti Nagar,🔝 9953056974 🔝 escort ServiceHot Sexy call girls in  Moti Nagar,🔝 9953056974 🔝 escort Service
Hot Sexy call girls in Moti Nagar,🔝 9953056974 🔝 escort Service
 
User Guide: Pulsar™ Weather Station (Columbia Weather Systems)
User Guide: Pulsar™ Weather Station (Columbia Weather Systems)User Guide: Pulsar™ Weather Station (Columbia Weather Systems)
User Guide: Pulsar™ Weather Station (Columbia Weather Systems)
 
Dubai Calls Girl Lisa O525547819 Lexi Call Girls In Dubai
Dubai Calls Girl Lisa O525547819 Lexi Call Girls In DubaiDubai Calls Girl Lisa O525547819 Lexi Call Girls In Dubai
Dubai Calls Girl Lisa O525547819 Lexi Call Girls In Dubai
 
STOPPED FLOW METHOD & APPLICATION MURUGAVENI B.pptx
STOPPED FLOW METHOD & APPLICATION MURUGAVENI B.pptxSTOPPED FLOW METHOD & APPLICATION MURUGAVENI B.pptx
STOPPED FLOW METHOD & APPLICATION MURUGAVENI B.pptx
 
Behavioral Disorder: Schizophrenia & it's Case Study.pdf
Behavioral Disorder: Schizophrenia & it's Case Study.pdfBehavioral Disorder: Schizophrenia & it's Case Study.pdf
Behavioral Disorder: Schizophrenia & it's Case Study.pdf
 
Harmful and Useful Microorganisms Presentation
Harmful and Useful Microorganisms PresentationHarmful and Useful Microorganisms Presentation
Harmful and Useful Microorganisms Presentation
 
Four Spheres of the Earth Presentation.ppt
Four Spheres of the Earth Presentation.pptFour Spheres of the Earth Presentation.ppt
Four Spheres of the Earth Presentation.ppt
 
Call Girls In Nihal Vihar Delhi ❤️8860477959 Looking Escorts In 24/7 Delhi NCR
Call Girls In Nihal Vihar Delhi ❤️8860477959 Looking Escorts In 24/7 Delhi NCRCall Girls In Nihal Vihar Delhi ❤️8860477959 Looking Escorts In 24/7 Delhi NCR
Call Girls In Nihal Vihar Delhi ❤️8860477959 Looking Escorts In 24/7 Delhi NCR
 
Pests of Bengal gram_Identification_Dr.UPR.pdf
Pests of Bengal gram_Identification_Dr.UPR.pdfPests of Bengal gram_Identification_Dr.UPR.pdf
Pests of Bengal gram_Identification_Dr.UPR.pdf
 
Citronella presentation SlideShare mani upadhyay
Citronella presentation SlideShare mani upadhyayCitronella presentation SlideShare mani upadhyay
Citronella presentation SlideShare mani upadhyay
 
Pests of jatropha_Bionomics_identification_Dr.UPR.pdf
Pests of jatropha_Bionomics_identification_Dr.UPR.pdfPests of jatropha_Bionomics_identification_Dr.UPR.pdf
Pests of jatropha_Bionomics_identification_Dr.UPR.pdf
 
《Queensland毕业文凭-昆士兰大学毕业证成绩单》
《Queensland毕业文凭-昆士兰大学毕业证成绩单》《Queensland毕业文凭-昆士兰大学毕业证成绩单》
《Queensland毕业文凭-昆士兰大学毕业证成绩单》
 
Best Call Girls In Sector 29 Gurgaon❤️8860477959 EscorTs Service In 24/7 Delh...
Best Call Girls In Sector 29 Gurgaon❤️8860477959 EscorTs Service In 24/7 Delh...Best Call Girls In Sector 29 Gurgaon❤️8860477959 EscorTs Service In 24/7 Delh...
Best Call Girls In Sector 29 Gurgaon❤️8860477959 EscorTs Service In 24/7 Delh...
 
preservation, maintanence and improvement of industrial organism.pptx
preservation, maintanence and improvement of industrial organism.pptxpreservation, maintanence and improvement of industrial organism.pptx
preservation, maintanence and improvement of industrial organism.pptx
 
BIOETHICS IN RECOMBINANT DNA TECHNOLOGY.
BIOETHICS IN RECOMBINANT DNA TECHNOLOGY.BIOETHICS IN RECOMBINANT DNA TECHNOLOGY.
BIOETHICS IN RECOMBINANT DNA TECHNOLOGY.
 
Microphone- characteristics,carbon microphone, dynamic microphone.pptx
Microphone- characteristics,carbon microphone, dynamic microphone.pptxMicrophone- characteristics,carbon microphone, dynamic microphone.pptx
Microphone- characteristics,carbon microphone, dynamic microphone.pptx
 
Environmental Biotechnology Topic:- Microbial Biosensor
Environmental Biotechnology Topic:- Microbial BiosensorEnvironmental Biotechnology Topic:- Microbial Biosensor
Environmental Biotechnology Topic:- Microbial Biosensor
 
Carbon Dioxide Capture and Storage (CSS)
Carbon Dioxide Capture and Storage (CSS)Carbon Dioxide Capture and Storage (CSS)
Carbon Dioxide Capture and Storage (CSS)
 

A Unifying Probabilistic Perspective for Spectral Dimensionality Reduction

  • 1. A Unifying Probabilistic Perspective for Spectral Dimensionality Reduction: Insights and New Models – Neil D. Lawrence Presenters: Sean Golliher and Derek Reimanis 1 / 21
  • 2. Claims and Contributions “Unifying Approach”: Views previous methods as less general that this approach and proves they are special cases Presents the Unified Algorithm: Maximum Entropy Unfolding (MEU) Improves Local Linear Embedding (LLE) and introduces ALLE (Acyclic Version) Introduces 3rd Algorithm DRILL. Estimating the dimensional structure from data (rather than K nearest neighbors) Mostly theoretical paper 2 / 21
  • 3. Features of Spectral Methods Spectral Methods: Low dimensional representations are derived from eigenvectors of specially constructed matrices. Methods Share a Common Approach: Compute nearest neighbors of input pattern Construct a weighted graph based on neighborhood relations Derive a matrix from the weighted graph Produce a representation of the matrix from the eigenvectors of the matrix Keep this mind when trying to understand unification 3 / 21
  • 4. Most Common Examples PCA: Technique that preserves the covariance of the data structure MDS Multidimensional Scaling: preserves inner products of input patterns. ex: xi − xj 2 ISOMAP: Preserves pairwise distances by measuring along a sub-manifold from the sample space. Variant of MDS which uses geodesics along sub-manifold geodesic: curve whose tangent vectors remain parallel if they are transported along it 4 / 21
  • 5. Local Linear Embedding (LLE) Tries to capture non-linear relationships by preserving local linear structure of data. Four Step Algorithm: Step 1: Compute K-nearest neighbors of each higher-dimensional input pattern xi and create directed graph whose edges indicate nearest neighbor relations. Step 2: Assign weights Wj to edges in graph. Each input pattern, and neighbors, can be viewed as samples from linear patch on a lower dimensional sub manifold. Find the “reconstruction weights” that provide a local linear fit of the k + 1 one points at each of the n points in the data set. Minimize error function: E(W|D) = x∈D x − y∈N(x) Wxyy 2 5 / 21
  • 6. Local Linear Embedding (LLE) Cont’d Step 3: After n local models are constructed keep weights fixed and find new projection. We turn this into a minimization problem similar to MDS E(Z|D) = x∈D zx − y∈N(x) Wxyzy 2 Put into a matrix Mxy that gives us a new objective function in matrix form (error function): E(Z|W ) = x,y Mxy(zx )T zy Find the m + 1 eigenvectors with the smallest eigenvectors and drop them. Giving us a new set of coordinates for new k-dimensional space. 6 / 21
  • 7. Maximum Entropy Unfolding (MEU) Kernel PCA idea: Apply “Kernal Trick” to transform to higher dimensional space. Then transform using covariance space as in PCA. Increases the features space rather than reducing the dimensions. This motived the development of Maximum Variance Unfolding (MVU) . MEU: Since entropy is related to variance they use entropy and obtain a probabilistic model. Derive a density over p(Y) directly (not over squared distances) by constraining the expected square inter-point distances dij of any two samples (yi ) and (yj ) 7 / 21
  • 8. Maximum Entropy Unfolding (MEU) Cont’d Observations are over squared distances they derive a density that gives rise to those distances p(Y). Uses KL divergence for calculating entropy between two distributions. However, we need two distributions: H = − p(Y)log p(Y) m(Y)dY Assume a very broad, spherical, Gaussian density that can be assumed to be small. Solving for Y the density that minimizes the KL divergence under the constraints on the expectations is: p(Y) = p j=1 |L+γI|1/2 τn/2 exp(−1/2yT j (L + γI))yj Where L is a special matrix where off diagonal elements contain information for D. Similar to previous examples. Key point is that this assumes independence of density across data features (multiplying densities) 8 / 21
  • 9. Maximum Entropy Unfolding (MEU) Cont’d GMRF: (finite-dimensional) random vector following a multivariate normal (or Gaussian) distribution. Since GRFs show provide an alternative approach to reducing number of parameters in covariance matrix they use GRF approach which is common on spatial methods. Independence in their model is being expressed over data features instead of data points. This assumes features are i.i.d For their model the number of parameters does not increase with the number of data. As number of features increases their is a clear “blessing of dimensionality”. 9 / 21
  • 10. Maximum Entropy Unfolding (MEU) Cont’d With the GRF representation they show that with some manipulation they can compute < dij >from the covariance matrix: < dij >= p/2(ki,j − 2ki,j + ki,j ) This is of the same form as distances and similarities in Kernel PCA methods 10 / 21
  • 11. MEU and LLE They show that LLE is approximating maximum likelihood in their MEU model Finding estimates of the parameters that maximize the probability of observing the data that we have Thus the claim of a “unifying model”. LLE is special case of MEU. Also showed that using pseudo-likelihood in the MEU model reduces to the generalization they presented in equation (9) for LLE. Using fact that join probability density of GRF can be represented as a factorization of the cliques of the graph. Pseudo-likelihood reduces the multiplication of the of the distribution presented earlier to the objective matrix for LLE. 11 / 21
  • 12. Acyclic Locally Linear Embedding (ALLE) Pseudo-likelihood approximation to the joint probability distribution of a collection of random variables. Thus the claim of a “unifying model”. LLE is special case of MEU. Also showed that using pseudo-likelihood in the MEU model reduces to the generalization they presented in equation (9) for LLE. If they force their matrix M to be lower triangular they that the covariance of the try log-likelihood of y log p(Y) factors into n independent regression functions. This approach requires an ordering of the data because of the restriction for j > i in the matrix M 12 / 21
  • 13. Dimensionality reduction through Regularization of the Inverse covariance in the Log Likelihood (DRILL) Optimization technique that applies L-1 regularization of dependencies to estimate graph structure How does this compare to the nearest-neighbor approaches? E(Λ) = − log p(Y ) + i<j λi,j 1 Maximize E(Λ) through Least Absolute Shrinkage and Selection Operator (LASSO) regression 13 / 21
  • 14. Gaussian process latent variable model (GP-LVM) likelihood Process which maps data from a latent space to a non-latent data space where the locale of points is determined by maximizing a Gaussian likelihood. This process allows us to map from a low dimensional space Rd (latent space) to a high dimensional space RD (data space) GP-LVM scoring Define hyperparameter θ = (σ, θrbf , θnoise) lml = max σ,θrbf ,θnoise log p(Y |X, θ) Find hyperparameter values through gradient descent 14 / 21
  • 19. GP-LVM scores for experiments 19 / 21
  • 20. Structural learning for DRILL 20 / 21
  • 21. Discussion Questions Is there an equivalent “no free lunch theorem” for dimensionality reduction? Why or why not? Were you convinced that this is a unifying approach? If this is a unifying approach why is MEU the worst performer? Does a theoretical paper such as this need more experiments? Is the GP-LVM score a good performance measure for comparing these methods? Are the experiments themselves introducing any bias in the results? Why do you think ALLE outperforms MEU? Is the “blessing of dimensionality” a valid claim? 21 / 21