SlideShare una empresa de Scribd logo
1 de 42
Accelerating C ollapsed  V ariational  B ayesian Inference   for L atent  D irichlet  A llocation  with Nvidia  CUDA  Compatible Devices Tomonari MASADA 正田 備也 Nagasaki University [email_address]
Overview ,[object Object],[object Object],[object Object],[object Object],Tomonari MASADA (IEA-AIE 2009)
LDA (latent Dirichlet allocation )
latent Dirichlet allocation [Blei et al. 02] ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],Tomonari MASADA (IEA-AIE 2009)
Dir ( β ) topic = word multinomial Tomonari MASADA (IEA-AIE 2009) symmetric Dirichlet prior v 1 v 2 v 3 v 4 t 1 φ 11 φ 12 φ 13 φ 14 v 1 v 2 v 3 v 4 t 2 φ 21 φ 22 φ 23 φ 24 v 1 v 2 v 3 v 4 t 3 φ 31 φ 32 φ 33 φ 34
document Dir ( α ) per-doc topic multinomial Tomonari MASADA (IEA-AIE 2009) symmetric Dirichlet prior θ j 1 θ j 2 θ j 3
v 3 v 1 v 3 v 2 v 2 θ j 1 θ j 2 θ j 3 Tomonari MASADA (IEA-AIE 2009) v 1 v 2 v 3 v 4 t 3 φ 31 φ 32 φ 33 φ 34 v 1 v 2 v 3 v 4 t 2 φ 21 φ 22 φ 23 φ 24 v 1 v 2 v 3 v 4 t 1 φ 11 φ 12 φ 13 φ 14
posterior distribution ,[object Object],[object Object],[object Object],[object Object],unknown known Tomonari MASADA (IEA-AIE 2009)
Inference methods for LDA ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],Tomonari MASADA (IEA-AIE 2009)
・・・ Tomonari MASADA (IEA-AIE 2009) vote γ 111 γ 112 γ 113 party γ 12 1 γ 122 γ 123 prime γ 131 γ 132 γ 133 ・・・ stock γ 241 γ 242 γ 243 ratio γ 251 γ 252 γ 253 prime γ 231 γ 232 γ 233 ・・・ party γ 321 γ 322 γ 323 celeb γ 361 γ 362 γ 363 prime γ 331 γ 332 γ 333 ・・・
Interpretation of  γ jwk   ,[object Object],[object Object],[object Object],[object Object],Tomonari MASADA (IEA-AIE 2009)
Algorithm of CVB ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],Update  γ jwk j : doc id w : word id k : topic id Tomonari MASADA (IEA-AIE 2009) O ( MK ) time M  : # of unique doc-word pairs K   : # of topics
Updating posterior parameters ,[object Object],[object Object],[object Object],[object Object],[object Object],Tomonari MASADA (IEA-AIE 2009) j : doc id w : word id k : topic id
Approximation by Gaussian ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],Tomonari MASADA (IEA-AIE 2009) j : doc id w : word id k : topic id
O( JK ) size O( KW ) size O( K ) size O( MK ) size Tomonari MASADA (IEA-AIE 2009) j : doc id w : word id k : topic id E[ n jk ] Var[ n jk ] E[ n kw ] Var[ n kw ] E[ n k ] Var[ n k ] γ jwk
[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],Details of CVB for LDA Update another two types of E[]s and Var[]s in a similar manner. Tomonari MASADA (IEA-AIE 2009) j : doc id w : word id k : topic id
Parallelization of CVB for LDA “ as many threads as topics”
Parallelization of CVB ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],Update  γ jwk Tomonari MASADA (IEA-AIE 2009) conventional parallelization proposed parallelization
[object Object],[object Object],[object Object],[object Object],[object Object],Strategy: “different topics for different threads” Tomonari MASADA (IEA-AIE 2009) γ jw 1   +  γ jw 2   +  ・・・  +  γ jwK  = 1 Normalization is required! O ( MK )     O ( M log K )
Reduction for normalization O (log K ) Tomonari MASADA (IEA-AIE 2009)
Related Work
A new approach (not a parallelization) ,[object Object],[object Object],[object Object],[object Object],[object Object],Tomonari MASADA (IEA-AIE 2009)
Accelerating CVB for LDA by GPGPU
Nvidia CUDA   (C ompute  U nified  D evice  A rchitecture )  Grid Device Memory documents topics Shared Memory Registers Thread Registers Thread Block Shared Memory Registers Thread Registers Thread Block
Device memory access latency Grid Device Memory 16KB Shared Memory Registers Thread Registers Thread Block Shared Memory Registers Thread Registers Thread Block
Data transfer latency Host Memory Transfer one large block instead of many smaller ones! Grid Device Memory Shared Memory Registers Thread Registers Thread Block Shared Memory Registers Thread Registers Thread Block
O( JK ) size O( KW ) size O( K ) size parameters of approximated posterior Tomonari MASADA (IEA-AIE 2009) j : doc id w : word id k : topic id E[ n jk ] Var[ n jk ] E[ n kw ] Var[ n kw ] E[ n k ] Var[ n k ] γ jwk
Where to store? ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],registers shared memory (for summation) registers device memory Tomonari MASADA (IEA-AIE 2009)
write conflicts O( KW ) size Tomonari MASADA (IEA-AIE 2009) γ j’wk E[ n kw ] Var[ n kw ] γ jwk
Experiments
Text mining ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],Tomonari MASADA (IEA-AIE 2009)
16 topics 64 iterations on CPU 64 iterations on GPU
64 iterations on CPU 64 iterations on GPU 32 topics
64 topics 64 iterations on CPU 64 iterations on GPU
Image mining ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],Tomonari MASADA (IEA-AIE 2009)
Image mining ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],Tomonari MASADA (IEA-AIE 2009)
Tomonari MASADA (IEA-AIE 2009)
Image mining (2 nd  trial) ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],Tomonari MASADA (IEA-AIE 2009)
Summary
discussions ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],Tomonari MASADA (IEA-AIE 2009)
Future work ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],Tomonari MASADA (IEA-AIE 2009)
Thank you for your attention! 非常感謝 !!!

Más contenido relacionado

La actualidad más candente

EuroPython 2017 - PyData - Deep Learning your Broadband Network @ HOME
EuroPython 2017 - PyData - Deep Learning your Broadband Network @ HOMEEuroPython 2017 - PyData - Deep Learning your Broadband Network @ HOME
EuroPython 2017 - PyData - Deep Learning your Broadband Network @ HOMEHONGJOO LEE
 
Speaker Diarization
Speaker DiarizationSpeaker Diarization
Speaker DiarizationHONGJOO LEE
 
High-Performance Approach to String Similarity using Most Frequent K Characters
High-Performance Approach to String Similarity using Most Frequent K CharactersHigh-Performance Approach to String Similarity using Most Frequent K Characters
High-Performance Approach to String Similarity using Most Frequent K CharactersHolistic Benchmarking of Big Linked Data
 
Querying Temporal Databases via OWL 2 QL
Querying Temporal Databases via OWL 2 QLQuerying Temporal Databases via OWL 2 QL
Querying Temporal Databases via OWL 2 QLSzymon Klarman
 
Information in the Weights
Information in the WeightsInformation in the Weights
Information in the WeightsMark Chang
 
Mining Frequent Closed Graphs on Evolving Data Streams
Mining Frequent Closed Graphs on Evolving Data StreamsMining Frequent Closed Graphs on Evolving Data Streams
Mining Frequent Closed Graphs on Evolving Data StreamsAlbert Bifet
 
Information in the Weights
Information in the WeightsInformation in the Weights
Information in the WeightsMark Chang
 
Parallel Implementation of K Means Clustering on CUDA
Parallel Implementation of K Means Clustering on CUDAParallel Implementation of K Means Clustering on CUDA
Parallel Implementation of K Means Clustering on CUDAprithan
 
A Short Course in Data Stream Mining
A Short Course in Data Stream MiningA Short Course in Data Stream Mining
A Short Course in Data Stream MiningAlbert Bifet
 
PyTorch 튜토리얼 (Touch to PyTorch)
PyTorch 튜토리얼 (Touch to PyTorch)PyTorch 튜토리얼 (Touch to PyTorch)
PyTorch 튜토리얼 (Touch to PyTorch)Hansol Kang
 
Large scale logistic regression and linear support vector machines using spark
Large scale logistic regression and linear support vector machines using sparkLarge scale logistic regression and linear support vector machines using spark
Large scale logistic regression and linear support vector machines using sparkMila, Université de Montréal
 
A Signature Scheme as Secure as the Diffie Hellman Problem
A Signature Scheme as Secure as the Diffie Hellman ProblemA Signature Scheme as Secure as the Diffie Hellman Problem
A Signature Scheme as Secure as the Diffie Hellman Problemvsubhashini
 
Gpu workshop cluster universe: scripting cuda
Gpu workshop cluster universe: scripting cudaGpu workshop cluster universe: scripting cuda
Gpu workshop cluster universe: scripting cudaFerdinand Jamitzky
 
GPU Accelerated Domain Decomposition
GPU Accelerated Domain DecompositionGPU Accelerated Domain Decomposition
GPU Accelerated Domain DecompositionRichard Southern
 
쉽게 설명하는 GAN (What is this? Gum? It's GAN.)
쉽게 설명하는 GAN (What is this? Gum? It's GAN.)쉽게 설명하는 GAN (What is this? Gum? It's GAN.)
쉽게 설명하는 GAN (What is this? Gum? It's GAN.)Hansol Kang
 
A TRAINING METHOD USING
 DNN-GUIDED LAYERWISE PRETRAINING
 FOR DEEP GAUSSIAN ...
A TRAINING METHOD USING
 DNN-GUIDED LAYERWISE PRETRAINING
 FOR DEEP GAUSSIAN ...A TRAINING METHOD USING
 DNN-GUIDED LAYERWISE PRETRAINING
 FOR DEEP GAUSSIAN ...
A TRAINING METHOD USING
 DNN-GUIDED LAYERWISE PRETRAINING
 FOR DEEP GAUSSIAN ...Tomoki Koriyama
 
LSGAN - SIMPle(Simple Idea Meaningful Performance Level up)
LSGAN - SIMPle(Simple Idea Meaningful Performance Level up)LSGAN - SIMPle(Simple Idea Meaningful Performance Level up)
LSGAN - SIMPle(Simple Idea Meaningful Performance Level up)Hansol Kang
 
PAC Bayesian for Deep Learning
PAC Bayesian for Deep LearningPAC Bayesian for Deep Learning
PAC Bayesian for Deep LearningMark Chang
 

La actualidad más candente (20)

Europy17_dibernardo
Europy17_dibernardoEuropy17_dibernardo
Europy17_dibernardo
 
EuroPython 2017 - PyData - Deep Learning your Broadband Network @ HOME
EuroPython 2017 - PyData - Deep Learning your Broadband Network @ HOMEEuroPython 2017 - PyData - Deep Learning your Broadband Network @ HOME
EuroPython 2017 - PyData - Deep Learning your Broadband Network @ HOME
 
Speaker Diarization
Speaker DiarizationSpeaker Diarization
Speaker Diarization
 
High-Performance Approach to String Similarity using Most Frequent K Characters
High-Performance Approach to String Similarity using Most Frequent K CharactersHigh-Performance Approach to String Similarity using Most Frequent K Characters
High-Performance Approach to String Similarity using Most Frequent K Characters
 
Querying Temporal Databases via OWL 2 QL
Querying Temporal Databases via OWL 2 QLQuerying Temporal Databases via OWL 2 QL
Querying Temporal Databases via OWL 2 QL
 
Information in the Weights
Information in the WeightsInformation in the Weights
Information in the Weights
 
Mining Frequent Closed Graphs on Evolving Data Streams
Mining Frequent Closed Graphs on Evolving Data StreamsMining Frequent Closed Graphs on Evolving Data Streams
Mining Frequent Closed Graphs on Evolving Data Streams
 
Lec 5-nn-slides
Lec 5-nn-slidesLec 5-nn-slides
Lec 5-nn-slides
 
Information in the Weights
Information in the WeightsInformation in the Weights
Information in the Weights
 
Parallel Implementation of K Means Clustering on CUDA
Parallel Implementation of K Means Clustering on CUDAParallel Implementation of K Means Clustering on CUDA
Parallel Implementation of K Means Clustering on CUDA
 
A Short Course in Data Stream Mining
A Short Course in Data Stream MiningA Short Course in Data Stream Mining
A Short Course in Data Stream Mining
 
PyTorch 튜토리얼 (Touch to PyTorch)
PyTorch 튜토리얼 (Touch to PyTorch)PyTorch 튜토리얼 (Touch to PyTorch)
PyTorch 튜토리얼 (Touch to PyTorch)
 
Large scale logistic regression and linear support vector machines using spark
Large scale logistic regression and linear support vector machines using sparkLarge scale logistic regression and linear support vector machines using spark
Large scale logistic regression and linear support vector machines using spark
 
A Signature Scheme as Secure as the Diffie Hellman Problem
A Signature Scheme as Secure as the Diffie Hellman ProblemA Signature Scheme as Secure as the Diffie Hellman Problem
A Signature Scheme as Secure as the Diffie Hellman Problem
 
Gpu workshop cluster universe: scripting cuda
Gpu workshop cluster universe: scripting cudaGpu workshop cluster universe: scripting cuda
Gpu workshop cluster universe: scripting cuda
 
GPU Accelerated Domain Decomposition
GPU Accelerated Domain DecompositionGPU Accelerated Domain Decomposition
GPU Accelerated Domain Decomposition
 
쉽게 설명하는 GAN (What is this? Gum? It's GAN.)
쉽게 설명하는 GAN (What is this? Gum? It's GAN.)쉽게 설명하는 GAN (What is this? Gum? It's GAN.)
쉽게 설명하는 GAN (What is this? Gum? It's GAN.)
 
A TRAINING METHOD USING
 DNN-GUIDED LAYERWISE PRETRAINING
 FOR DEEP GAUSSIAN ...
A TRAINING METHOD USING
 DNN-GUIDED LAYERWISE PRETRAINING
 FOR DEEP GAUSSIAN ...A TRAINING METHOD USING
 DNN-GUIDED LAYERWISE PRETRAINING
 FOR DEEP GAUSSIAN ...
A TRAINING METHOD USING
 DNN-GUIDED LAYERWISE PRETRAINING
 FOR DEEP GAUSSIAN ...
 
LSGAN - SIMPle(Simple Idea Meaningful Performance Level up)
LSGAN - SIMPle(Simple Idea Meaningful Performance Level up)LSGAN - SIMPle(Simple Idea Meaningful Performance Level up)
LSGAN - SIMPle(Simple Idea Meaningful Performance Level up)
 
PAC Bayesian for Deep Learning
PAC Bayesian for Deep LearningPAC Bayesian for Deep Learning
PAC Bayesian for Deep Learning
 

Destacado

SF Big Analytics & SF Machine Learning Meetup: Machine Learning at the Limit ...
SF Big Analytics & SF Machine Learning Meetup: Machine Learning at the Limit ...SF Big Analytics & SF Machine Learning Meetup: Machine Learning at the Limit ...
SF Big Analytics & SF Machine Learning Meetup: Machine Learning at the Limit ...Chester Chen
 
WSDM2014
WSDM2014WSDM2014
WSDM2014Jun Yu
 
Interactive Latent Dirichlet Allocation
Interactive Latent Dirichlet AllocationInteractive Latent Dirichlet Allocation
Interactive Latent Dirichlet AllocationQuentin Pleplé
 
Goal-based Recommendation utilizing Latent Dirichlet Allocation
Goal-based Recommendation utilizing Latent Dirichlet AllocationGoal-based Recommendation utilizing Latent Dirichlet Allocation
Goal-based Recommendation utilizing Latent Dirichlet AllocationSebastien Louvigne
 
A Note on Expectation-Propagation for Latent Dirichlet Allocation
A Note on Expectation-Propagation for Latent Dirichlet AllocationA Note on Expectation-Propagation for Latent Dirichlet Allocation
A Note on Expectation-Propagation for Latent Dirichlet AllocationTomonari Masada
 
What's new in scikit-learn 0.17
What's new in scikit-learn 0.17What's new in scikit-learn 0.17
What's new in scikit-learn 0.17Andreas Mueller
 
Latent dirichletallocation presentation
Latent dirichletallocation presentationLatent dirichletallocation presentation
Latent dirichletallocation presentationSoojung Hong
 
Visualzing Topic Models
Visualzing Topic ModelsVisualzing Topic Models
Visualzing Topic ModelsTuri, Inc.
 
A Simple Stochastic Gradient Variational Bayes for the Correlated Topic Model
A Simple Stochastic Gradient Variational Bayes for the Correlated Topic ModelA Simple Stochastic Gradient Variational Bayes for the Correlated Topic Model
A Simple Stochastic Gradient Variational Bayes for the Correlated Topic ModelTomonari Masada
 
Latent Dirichlet Allocation
Latent Dirichlet AllocationLatent Dirichlet Allocation
Latent Dirichlet AllocationMarco Righini
 
A Simple Stochastic Gradient Variational Bayes for Latent Dirichlet Allocation
A Simple Stochastic Gradient Variational Bayes for Latent Dirichlet AllocationA Simple Stochastic Gradient Variational Bayes for Latent Dirichlet Allocation
A Simple Stochastic Gradient Variational Bayes for Latent Dirichlet AllocationTomonari Masada
 
Blei ngjordan2003
Blei ngjordan2003Blei ngjordan2003
Blei ngjordan2003Ajay Ohri
 
Latent Dirichlet Allocation as a Twitter Hashtag Recommendation System
Latent Dirichlet Allocation as a Twitter Hashtag Recommendation SystemLatent Dirichlet Allocation as a Twitter Hashtag Recommendation System
Latent Dirichlet Allocation as a Twitter Hashtag Recommendation SystemShailly Saxena
 
LDA Beginner's Tutorial
LDA Beginner's TutorialLDA Beginner's Tutorial
LDA Beginner's TutorialWayne Lee
 

Destacado (18)

SF Big Analytics & SF Machine Learning Meetup: Machine Learning at the Limit ...
SF Big Analytics & SF Machine Learning Meetup: Machine Learning at the Limit ...SF Big Analytics & SF Machine Learning Meetup: Machine Learning at the Limit ...
SF Big Analytics & SF Machine Learning Meetup: Machine Learning at the Limit ...
 
WSDM2014
WSDM2014WSDM2014
WSDM2014
 
Interactive Latent Dirichlet Allocation
Interactive Latent Dirichlet AllocationInteractive Latent Dirichlet Allocation
Interactive Latent Dirichlet Allocation
 
Goal-based Recommendation utilizing Latent Dirichlet Allocation
Goal-based Recommendation utilizing Latent Dirichlet AllocationGoal-based Recommendation utilizing Latent Dirichlet Allocation
Goal-based Recommendation utilizing Latent Dirichlet Allocation
 
A Note on Expectation-Propagation for Latent Dirichlet Allocation
A Note on Expectation-Propagation for Latent Dirichlet AllocationA Note on Expectation-Propagation for Latent Dirichlet Allocation
A Note on Expectation-Propagation for Latent Dirichlet Allocation
 
What's new in scikit-learn 0.17
What's new in scikit-learn 0.17What's new in scikit-learn 0.17
What's new in scikit-learn 0.17
 
Latent dirichletallocation presentation
Latent dirichletallocation presentationLatent dirichletallocation presentation
Latent dirichletallocation presentation
 
Visualzing Topic Models
Visualzing Topic ModelsVisualzing Topic Models
Visualzing Topic Models
 
Topic Modeling
Topic ModelingTopic Modeling
Topic Modeling
 
A Simple Stochastic Gradient Variational Bayes for the Correlated Topic Model
A Simple Stochastic Gradient Variational Bayes for the Correlated Topic ModelA Simple Stochastic Gradient Variational Bayes for the Correlated Topic Model
A Simple Stochastic Gradient Variational Bayes for the Correlated Topic Model
 
Latent Dirichlet Allocation
Latent Dirichlet AllocationLatent Dirichlet Allocation
Latent Dirichlet Allocation
 
A Simple Stochastic Gradient Variational Bayes for Latent Dirichlet Allocation
A Simple Stochastic Gradient Variational Bayes for Latent Dirichlet AllocationA Simple Stochastic Gradient Variational Bayes for Latent Dirichlet Allocation
A Simple Stochastic Gradient Variational Bayes for Latent Dirichlet Allocation
 
Blei ngjordan2003
Blei ngjordan2003Blei ngjordan2003
Blei ngjordan2003
 
C4.5
C4.5C4.5
C4.5
 
Topic Models
Topic ModelsTopic Models
Topic Models
 
LDA入門
LDA入門LDA入門
LDA入門
 
Latent Dirichlet Allocation as a Twitter Hashtag Recommendation System
Latent Dirichlet Allocation as a Twitter Hashtag Recommendation SystemLatent Dirichlet Allocation as a Twitter Hashtag Recommendation System
Latent Dirichlet Allocation as a Twitter Hashtag Recommendation System
 
LDA Beginner's Tutorial
LDA Beginner's TutorialLDA Beginner's Tutorial
LDA Beginner's Tutorial
 

Similar a Accelerating Collapsed Variational Bayesian Inference for Latent Dirichlet Allocation with Nvidia CUDA compatible devices

Steering Time-Dependent Estimation of Posteriors with Hyperparameter Indexing...
Steering Time-Dependent Estimation of Posteriors with Hyperparameter Indexing...Steering Time-Dependent Estimation of Posteriors with Hyperparameter Indexing...
Steering Time-Dependent Estimation of Posteriors with Hyperparameter Indexing...Tomonari Masada
 
NÜWA: Visual Synthesis Pre-training for Neural visUal World creAtion
NÜWA: Visual Synthesis Pre-training for Neural visUal World creAtionNÜWA: Visual Synthesis Pre-training for Neural visUal World creAtion
NÜWA: Visual Synthesis Pre-training for Neural visUal World creAtionKai Katsumata
 
11.iterative idma receivers with random and tree based interleavers
11.iterative idma receivers with random and tree based interleavers11.iterative idma receivers with random and tree based interleavers
11.iterative idma receivers with random and tree based interleaversAlexander Decker
 
11.iterative idma receivers with random and tree based interleavers
11.iterative idma receivers with random and tree based interleavers11.iterative idma receivers with random and tree based interleavers
11.iterative idma receivers with random and tree based interleaversAlexander Decker
 
Iterative idma receivers with random and tree based interleavers
Iterative idma receivers with random and tree based interleaversIterative idma receivers with random and tree based interleavers
Iterative idma receivers with random and tree based interleaversAlexander Decker
 
Fine Grained Complexity of Rainbow Coloring and its Variants
Fine Grained Complexity of Rainbow Coloring and its VariantsFine Grained Complexity of Rainbow Coloring and its Variants
Fine Grained Complexity of Rainbow Coloring and its VariantsAkankshaAgrawal55
 
Kernel for Chordal Vertex Deletion
Kernel for Chordal Vertex DeletionKernel for Chordal Vertex Deletion
Kernel for Chordal Vertex DeletionAkankshaAgrawal55
 
On the Parameterized Complexity of Simultaneous Deletion Problems
On the Parameterized Complexity of Simultaneous Deletion ProblemsOn the Parameterized Complexity of Simultaneous Deletion Problems
On the Parameterized Complexity of Simultaneous Deletion ProblemsAkankshaAgrawal55
 
DAW: Duplicate-AWare Federated Query Processing over the Web of Data
DAW: Duplicate-AWare Federated Query Processing over the Web of DataDAW: Duplicate-AWare Federated Query Processing over the Web of Data
DAW: Duplicate-AWare Federated Query Processing over the Web of DataMuhammad Saleem
 
A Fast Near Optimal Vertex Cover Algorithm (NOVCA)
A Fast Near Optimal Vertex Cover Algorithm (NOVCA)A Fast Near Optimal Vertex Cover Algorithm (NOVCA)
A Fast Near Optimal Vertex Cover Algorithm (NOVCA)Waqas Tariq
 
Picmet15sasaki20150805.ppt
Picmet15sasaki20150805.pptPicmet15sasaki20150805.ppt
Picmet15sasaki20150805.pptHajime Sasaki
 
Split Contraction: The Untold Story
Split Contraction: The Untold StorySplit Contraction: The Untold Story
Split Contraction: The Untold StoryAkankshaAgrawal55
 
Graph Modification: Beyond the known Boundaries
Graph Modification: Beyond the known BoundariesGraph Modification: Beyond the known Boundaries
Graph Modification: Beyond the known BoundariesAkankshaAgrawal55
 
MongoDB, Hadoop and humongous data - MongoSV 2012
MongoDB, Hadoop and humongous data - MongoSV 2012MongoDB, Hadoop and humongous data - MongoSV 2012
MongoDB, Hadoop and humongous data - MongoSV 2012Steven Francia
 
Topic modeling with Poisson factorization (2)
Topic modeling with Poisson factorization (2)Topic modeling with Poisson factorization (2)
Topic modeling with Poisson factorization (2)Tomonari Masada
 
Multiscale methods for graphene based nanocomposites
Multiscale methods for graphene based nanocompositesMultiscale methods for graphene based nanocomposites
Multiscale methods for graphene based nanocompositesUniversity of Glasgow
 

Similar a Accelerating Collapsed Variational Bayesian Inference for Latent Dirichlet Allocation with Nvidia CUDA compatible devices (20)

Steering Time-Dependent Estimation of Posteriors with Hyperparameter Indexing...
Steering Time-Dependent Estimation of Posteriors with Hyperparameter Indexing...Steering Time-Dependent Estimation of Posteriors with Hyperparameter Indexing...
Steering Time-Dependent Estimation of Posteriors with Hyperparameter Indexing...
 
NÜWA: Visual Synthesis Pre-training for Neural visUal World creAtion
NÜWA: Visual Synthesis Pre-training for Neural visUal World creAtionNÜWA: Visual Synthesis Pre-training for Neural visUal World creAtion
NÜWA: Visual Synthesis Pre-training for Neural visUal World creAtion
 
11.iterative idma receivers with random and tree based interleavers
11.iterative idma receivers with random and tree based interleavers11.iterative idma receivers with random and tree based interleavers
11.iterative idma receivers with random and tree based interleavers
 
11.iterative idma receivers with random and tree based interleavers
11.iterative idma receivers with random and tree based interleavers11.iterative idma receivers with random and tree based interleavers
11.iterative idma receivers with random and tree based interleavers
 
Iterative idma receivers with random and tree based interleavers
Iterative idma receivers with random and tree based interleaversIterative idma receivers with random and tree based interleavers
Iterative idma receivers with random and tree based interleavers
 
Fine Grained Complexity of Rainbow Coloring and its Variants
Fine Grained Complexity of Rainbow Coloring and its VariantsFine Grained Complexity of Rainbow Coloring and its Variants
Fine Grained Complexity of Rainbow Coloring and its Variants
 
Kernel for Chordal Vertex Deletion
Kernel for Chordal Vertex DeletionKernel for Chordal Vertex Deletion
Kernel for Chordal Vertex Deletion
 
On the Parameterized Complexity of Simultaneous Deletion Problems
On the Parameterized Complexity of Simultaneous Deletion ProblemsOn the Parameterized Complexity of Simultaneous Deletion Problems
On the Parameterized Complexity of Simultaneous Deletion Problems
 
DAW: Duplicate-AWare Federated Query Processing over the Web of Data
DAW: Duplicate-AWare Federated Query Processing over the Web of DataDAW: Duplicate-AWare Federated Query Processing over the Web of Data
DAW: Duplicate-AWare Federated Query Processing over the Web of Data
 
A Fast Near Optimal Vertex Cover Algorithm (NOVCA)
A Fast Near Optimal Vertex Cover Algorithm (NOVCA)A Fast Near Optimal Vertex Cover Algorithm (NOVCA)
A Fast Near Optimal Vertex Cover Algorithm (NOVCA)
 
Poisson factorization
Poisson factorizationPoisson factorization
Poisson factorization
 
Picmet15sasaki20150805.ppt
Picmet15sasaki20150805.pptPicmet15sasaki20150805.ppt
Picmet15sasaki20150805.ppt
 
Split Contraction: The Untold Story
Split Contraction: The Untold StorySplit Contraction: The Untold Story
Split Contraction: The Untold Story
 
Graph Modification: Beyond the known Boundaries
Graph Modification: Beyond the known BoundariesGraph Modification: Beyond the known Boundaries
Graph Modification: Beyond the known Boundaries
 
MongoDB, Hadoop and humongous data - MongoSV 2012
MongoDB, Hadoop and humongous data - MongoSV 2012MongoDB, Hadoop and humongous data - MongoSV 2012
MongoDB, Hadoop and humongous data - MongoSV 2012
 
Topic modeling with Poisson factorization (2)
Topic modeling with Poisson factorization (2)Topic modeling with Poisson factorization (2)
Topic modeling with Poisson factorization (2)
 
Biochip
BiochipBiochip
Biochip
 
Multiscale methods for graphene based nanocomposites
Multiscale methods for graphene based nanocompositesMultiscale methods for graphene based nanocomposites
Multiscale methods for graphene based nanocomposites
 
Polynomial Matrix Decompositions
Polynomial Matrix DecompositionsPolynomial Matrix Decompositions
Polynomial Matrix Decompositions
 
discrete-hmm
discrete-hmmdiscrete-hmm
discrete-hmm
 

Más de Tomonari Masada

Learning Latent Space Energy Based Prior Modelの解説
Learning Latent Space Energy Based Prior Modelの解説Learning Latent Space Energy Based Prior Modelの解説
Learning Latent Space Energy Based Prior Modelの解説Tomonari Masada
 
Denoising Diffusion Probabilistic Modelsの重要な式の解説
Denoising Diffusion Probabilistic Modelsの重要な式の解説Denoising Diffusion Probabilistic Modelsの重要な式の解説
Denoising Diffusion Probabilistic Modelsの重要な式の解説Tomonari Masada
 
Context-dependent Token-wise Variational Autoencoder for Topic Modeling
Context-dependent Token-wise Variational Autoencoder for Topic ModelingContext-dependent Token-wise Variational Autoencoder for Topic Modeling
Context-dependent Token-wise Variational Autoencoder for Topic ModelingTomonari Masada
 
A note on the density of Gumbel-softmax
A note on the density of Gumbel-softmaxA note on the density of Gumbel-softmax
A note on the density of Gumbel-softmaxTomonari Masada
 
トピックモデルの基礎と応用
トピックモデルの基礎と応用トピックモデルの基礎と応用
トピックモデルの基礎と応用Tomonari Masada
 
Expectation propagation for latent Dirichlet allocation
Expectation propagation for latent Dirichlet allocationExpectation propagation for latent Dirichlet allocation
Expectation propagation for latent Dirichlet allocationTomonari Masada
 
Mini-batch Variational Inference for Time-Aware Topic Modeling
Mini-batch Variational Inference for Time-Aware Topic ModelingMini-batch Variational Inference for Time-Aware Topic Modeling
Mini-batch Variational Inference for Time-Aware Topic ModelingTomonari Masada
 
A note on variational inference for the univariate Gaussian
A note on variational inference for the univariate GaussianA note on variational inference for the univariate Gaussian
A note on variational inference for the univariate GaussianTomonari Masada
 
Document Modeling with Implicit Approximate Posterior Distributions
Document Modeling with Implicit Approximate Posterior DistributionsDocument Modeling with Implicit Approximate Posterior Distributions
Document Modeling with Implicit Approximate Posterior DistributionsTomonari Masada
 
LDA-Based Scoring of Sequences Generated by RNN for Automatic Tanka Composition
LDA-Based Scoring of Sequences Generated by RNN for Automatic Tanka CompositionLDA-Based Scoring of Sequences Generated by RNN for Automatic Tanka Composition
LDA-Based Scoring of Sequences Generated by RNN for Automatic Tanka CompositionTomonari Masada
 
A Note on Latent LSTM Allocation
A Note on Latent LSTM AllocationA Note on Latent LSTM Allocation
A Note on Latent LSTM AllocationTomonari Masada
 
Word count in Husserliana Volumes 1 to 28
Word count in Husserliana Volumes 1 to 28Word count in Husserliana Volumes 1 to 28
Word count in Husserliana Volumes 1 to 28Tomonari Masada
 
A Simple Stochastic Gradient Variational Bayes for Latent Dirichlet Allocation
A Simple Stochastic Gradient Variational Bayes for Latent Dirichlet AllocationA Simple Stochastic Gradient Variational Bayes for Latent Dirichlet Allocation
A Simple Stochastic Gradient Variational Bayes for Latent Dirichlet AllocationTomonari Masada
 
A derivation of the sampling formulas for An Entity-Topic Model for Entity Li...
A derivation of the sampling formulas for An Entity-Topic Model for Entity Li...A derivation of the sampling formulas for An Entity-Topic Model for Entity Li...
A derivation of the sampling formulas for An Entity-Topic Model for Entity Li...Tomonari Masada
 
A Note on BPTT for LSTM LM
A Note on BPTT for LSTM LMA Note on BPTT for LSTM LM
A Note on BPTT for LSTM LMTomonari Masada
 
The detailed derivation of the derivatives in Table 2 of Marginalized Denoisi...
The detailed derivation of the derivatives in Table 2 of Marginalized Denoisi...The detailed derivation of the derivatives in Table 2 of Marginalized Denoisi...
The detailed derivation of the derivatives in Table 2 of Marginalized Denoisi...Tomonari Masada
 
A Note on PCVB0 for HDP-LDA
A Note on PCVB0 for HDP-LDAA Note on PCVB0 for HDP-LDA
A Note on PCVB0 for HDP-LDATomonari Masada
 

Más de Tomonari Masada (20)

Learning Latent Space Energy Based Prior Modelの解説
Learning Latent Space Energy Based Prior Modelの解説Learning Latent Space Energy Based Prior Modelの解説
Learning Latent Space Energy Based Prior Modelの解説
 
Denoising Diffusion Probabilistic Modelsの重要な式の解説
Denoising Diffusion Probabilistic Modelsの重要な式の解説Denoising Diffusion Probabilistic Modelsの重要な式の解説
Denoising Diffusion Probabilistic Modelsの重要な式の解説
 
Context-dependent Token-wise Variational Autoencoder for Topic Modeling
Context-dependent Token-wise Variational Autoencoder for Topic ModelingContext-dependent Token-wise Variational Autoencoder for Topic Modeling
Context-dependent Token-wise Variational Autoencoder for Topic Modeling
 
A note on the density of Gumbel-softmax
A note on the density of Gumbel-softmaxA note on the density of Gumbel-softmax
A note on the density of Gumbel-softmax
 
トピックモデルの基礎と応用
トピックモデルの基礎と応用トピックモデルの基礎と応用
トピックモデルの基礎と応用
 
Expectation propagation for latent Dirichlet allocation
Expectation propagation for latent Dirichlet allocationExpectation propagation for latent Dirichlet allocation
Expectation propagation for latent Dirichlet allocation
 
Mini-batch Variational Inference for Time-Aware Topic Modeling
Mini-batch Variational Inference for Time-Aware Topic ModelingMini-batch Variational Inference for Time-Aware Topic Modeling
Mini-batch Variational Inference for Time-Aware Topic Modeling
 
A note on variational inference for the univariate Gaussian
A note on variational inference for the univariate GaussianA note on variational inference for the univariate Gaussian
A note on variational inference for the univariate Gaussian
 
Document Modeling with Implicit Approximate Posterior Distributions
Document Modeling with Implicit Approximate Posterior DistributionsDocument Modeling with Implicit Approximate Posterior Distributions
Document Modeling with Implicit Approximate Posterior Distributions
 
LDA-Based Scoring of Sequences Generated by RNN for Automatic Tanka Composition
LDA-Based Scoring of Sequences Generated by RNN for Automatic Tanka CompositionLDA-Based Scoring of Sequences Generated by RNN for Automatic Tanka Composition
LDA-Based Scoring of Sequences Generated by RNN for Automatic Tanka Composition
 
A Note on ZINB-VAE
A Note on ZINB-VAEA Note on ZINB-VAE
A Note on ZINB-VAE
 
A Note on Latent LSTM Allocation
A Note on Latent LSTM AllocationA Note on Latent LSTM Allocation
A Note on Latent LSTM Allocation
 
A Note on TopicRNN
A Note on TopicRNNA Note on TopicRNN
A Note on TopicRNN
 
Word count in Husserliana Volumes 1 to 28
Word count in Husserliana Volumes 1 to 28Word count in Husserliana Volumes 1 to 28
Word count in Husserliana Volumes 1 to 28
 
A Simple Stochastic Gradient Variational Bayes for Latent Dirichlet Allocation
A Simple Stochastic Gradient Variational Bayes for Latent Dirichlet AllocationA Simple Stochastic Gradient Variational Bayes for Latent Dirichlet Allocation
A Simple Stochastic Gradient Variational Bayes for Latent Dirichlet Allocation
 
FDSE2015
FDSE2015FDSE2015
FDSE2015
 
A derivation of the sampling formulas for An Entity-Topic Model for Entity Li...
A derivation of the sampling formulas for An Entity-Topic Model for Entity Li...A derivation of the sampling formulas for An Entity-Topic Model for Entity Li...
A derivation of the sampling formulas for An Entity-Topic Model for Entity Li...
 
A Note on BPTT for LSTM LM
A Note on BPTT for LSTM LMA Note on BPTT for LSTM LM
A Note on BPTT for LSTM LM
 
The detailed derivation of the derivatives in Table 2 of Marginalized Denoisi...
The detailed derivation of the derivatives in Table 2 of Marginalized Denoisi...The detailed derivation of the derivatives in Table 2 of Marginalized Denoisi...
The detailed derivation of the derivatives in Table 2 of Marginalized Denoisi...
 
A Note on PCVB0 for HDP-LDA
A Note on PCVB0 for HDP-LDAA Note on PCVB0 for HDP-LDA
A Note on PCVB0 for HDP-LDA
 

Accelerating Collapsed Variational Bayesian Inference for Latent Dirichlet Allocation with Nvidia CUDA compatible devices

  • 1. Accelerating C ollapsed V ariational B ayesian Inference for L atent D irichlet A llocation with Nvidia CUDA Compatible Devices Tomonari MASADA 正田 備也 Nagasaki University [email_address]
  • 2.
  • 3. LDA (latent Dirichlet allocation )
  • 4.
  • 5. Dir ( β ) topic = word multinomial Tomonari MASADA (IEA-AIE 2009) symmetric Dirichlet prior v 1 v 2 v 3 v 4 t 1 φ 11 φ 12 φ 13 φ 14 v 1 v 2 v 3 v 4 t 2 φ 21 φ 22 φ 23 φ 24 v 1 v 2 v 3 v 4 t 3 φ 31 φ 32 φ 33 φ 34
  • 6. document Dir ( α ) per-doc topic multinomial Tomonari MASADA (IEA-AIE 2009) symmetric Dirichlet prior θ j 1 θ j 2 θ j 3
  • 7. v 3 v 1 v 3 v 2 v 2 θ j 1 θ j 2 θ j 3 Tomonari MASADA (IEA-AIE 2009) v 1 v 2 v 3 v 4 t 3 φ 31 φ 32 φ 33 φ 34 v 1 v 2 v 3 v 4 t 2 φ 21 φ 22 φ 23 φ 24 v 1 v 2 v 3 v 4 t 1 φ 11 φ 12 φ 13 φ 14
  • 8.
  • 9.
  • 10. ・・・ Tomonari MASADA (IEA-AIE 2009) vote γ 111 γ 112 γ 113 party γ 12 1 γ 122 γ 123 prime γ 131 γ 132 γ 133 ・・・ stock γ 241 γ 242 γ 243 ratio γ 251 γ 252 γ 253 prime γ 231 γ 232 γ 233 ・・・ party γ 321 γ 322 γ 323 celeb γ 361 γ 362 γ 363 prime γ 331 γ 332 γ 333 ・・・
  • 11.
  • 12.
  • 13.
  • 14.
  • 15. O( JK ) size O( KW ) size O( K ) size O( MK ) size Tomonari MASADA (IEA-AIE 2009) j : doc id w : word id k : topic id E[ n jk ] Var[ n jk ] E[ n kw ] Var[ n kw ] E[ n k ] Var[ n k ] γ jwk
  • 16.
  • 17. Parallelization of CVB for LDA “ as many threads as topics”
  • 18.
  • 19.
  • 20. Reduction for normalization O (log K ) Tomonari MASADA (IEA-AIE 2009)
  • 22.
  • 23. Accelerating CVB for LDA by GPGPU
  • 24. Nvidia CUDA (C ompute U nified D evice A rchitecture ) Grid Device Memory documents topics Shared Memory Registers Thread Registers Thread Block Shared Memory Registers Thread Registers Thread Block
  • 25. Device memory access latency Grid Device Memory 16KB Shared Memory Registers Thread Registers Thread Block Shared Memory Registers Thread Registers Thread Block
  • 26. Data transfer latency Host Memory Transfer one large block instead of many smaller ones! Grid Device Memory Shared Memory Registers Thread Registers Thread Block Shared Memory Registers Thread Registers Thread Block
  • 27. O( JK ) size O( KW ) size O( K ) size parameters of approximated posterior Tomonari MASADA (IEA-AIE 2009) j : doc id w : word id k : topic id E[ n jk ] Var[ n jk ] E[ n kw ] Var[ n kw ] E[ n k ] Var[ n k ] γ jwk
  • 28.
  • 29. write conflicts O( KW ) size Tomonari MASADA (IEA-AIE 2009) γ j’wk E[ n kw ] Var[ n kw ] γ jwk
  • 31.
  • 32. 16 topics 64 iterations on CPU 64 iterations on GPU
  • 33. 64 iterations on CPU 64 iterations on GPU 32 topics
  • 34. 64 topics 64 iterations on CPU 64 iterations on GPU
  • 35.
  • 36.
  • 38.
  • 40.
  • 41.
  • 42. Thank you for your attention! 非常感謝 !!!