Filters








2,636 Hits in 8.2 sec

Sparse Nonparametric Topic Model for Transfer Learning

Ali Faisal, Jussi Gillberg, Jaakko Peltonen, Gayle Leen, Samuel Kaski
2012 The European Symposium on Artificial Neural Networks  
We introduce a Bayesian generative transfer learning model which represents similarity across document collections by sparse sharing of latent topics controlled by an Indian Buffet Process.  ...  With several data sets available from related sources, exploiting their similarities by transfer learning can improve models compared to modeling sources independently.  ...  New sparse nonparametric topic model for transfer learning We present a new hierarchical Bayesian multi-task (transfer learning) model which allows flexible sharing of low-strength and high-strength topics  ... 
dblp:conf/esann/FaisalGPLK12 fatcat:zhu6quglvrb5dhuqkltrtnwg4y

Transfer learning using a nonparametric sparse topic model

Ali Faisal, Jussi Gillberg, Gayle Leen, Jaakko Peltonen
2013 Neurocomputing  
We introduce a Bayesian generative transfer learning model which represents similarity across document collections by sparse sharing of latent topics controlled by an Indian Buffet Process.  ...  Unlike a prominent previous model, Hierarchical Dirichlet Process (HDP) based multi-task learning, our model decouples topic sharing probability from topic strength, making sharing of low-strength topics  ...  New sparse nonparametric topic model for transfer learning We present a new hierarchical Bayesian multi-task (transfer learning) model which allows flexible sharing of low-strength and high-strength topics  ... 
doi:10.1016/j.neucom.2012.12.038 fatcat:qnhlkfzdffhopj3c7jwkee332q

Sparse Parallel Training of Hierarchical Dirichlet Process Topic Models [article]

Alexander Terenin, Måns Magnusson, Leif Jonsson
2020 arXiv   pre-print
Based upon a representation of certain conditional distributions within an HDP, we propose a doubly sparse data-parallel sampler for the HDP topic model.  ...  In this work, we study data-parallel training for the hierarchical Dirichlet process (HDP) topic model.  ...  Nonparametric topic models are less straightforward to evaluate empirically than ordinary topic models.  ... 
arXiv:1906.02416v2 fatcat:rbsw4dud3rf7hmvj5sufd6uou4

Statistical Model Aggregation via Parameter Matching [article]

Mikhail Yurochkin, Mayank Agarwal, Soumya Ghosh, Kristjan Greenewald, Trong Nghia Hoang
2019 arXiv   pre-print
After verifying our approach on simulated data, we demonstrate its utility in aggregating Gaussian topic models, hierarchical Dirichlet process based hidden Markov models, and sparse Gaussian processes  ...  Exploiting tools from Bayesian nonparametrics, we develop a general meta-modeling framework that learns shared global latent structures by identifying correspondences among local model parameterizations  ...  Bayesian nonparametric federated learning of neural networks.  ... 
arXiv:1911.00218v1 fatcat:hoo254bj5reppmv4jlmx3q3vde

Dependent Indian Buffet Process-based Sparse Nonparametric Nonnegative Matrix Factorization [article]

Junyu Xuan and Jie Lu and Guangquan Zhang and Richard Yi Da Xu and Xiangfeng Luo
2015 arXiv   pre-print
The method has been widely used for unsupervised learning tasks, including recommender systems (rating matrix of users by items) and document clustering (weighting matrix of papers by keywords).  ...  This assumption makes them inflexible for many applications. In this paper, we propose a nonparametric NMF framework to mitigate this issue by using dependent Indian Buffet Processes (dIBP).  ...  Zimmer, Copula modeling: an introduc- [6] M. Heiler and C. Schnörr, “Learning sparse representations tion for practitioners. Now Publishers Inc, 2007.  ... 
arXiv:1507.03176v1 fatcat:jdt3mihx5fftng3n4nb2hympdq

Random Function Priors for Correlation Modeling [article]

Aonan Zhang, John Paisley
2019 arXiv   pre-print
Our model can be viewed as a generalized paintbox model Broderick13 using random functions, and can be learned efficiently with neural networks via amortized variational inference.  ...  In this paper, we introduce random function priors for Z_n for modeling correlations among its K dimensions Z_n1 through Z_nK, which we call population random measure embedding (PRME).  ...  Acknowledgements We thank Howard Karloff and Victor Veitch for their helpful comments during the early stage of this work.  ... 
arXiv:1905.03826v2 fatcat:zvmgegsfljcxna42ivehm7zpwu

A Survey on Bayesian Nonparametric Learning

Junyu Xuan, Jie Lu, Guangquan Zhang
2019 ACM Computing Surveys  
Here, BNL can not only learn the summarised topics in a set of documents but can also adapt the number of learned topics according to the documents in the set.  ...  The Bayesian nonparametric extensions of current machine learning algorithms or models have been reviewed as motivating examples for researchers who already have knowledge in machine learning.  ...  CONCLUDING REMARKS Bayesian nonparametric learning (BNL) is becoming a hot topic in machine learning due to its unique characteristics.  ... 
doi:10.1145/3291044 fatcat:aytdnsnrfvfnti5i64ne4icenu

Anomaly Feature Learning for Unsupervised Change Detection in Heterogeneous Images: A Deep Sparse Residual Model

Redha Touati, Max Mignotte, Mohamed Dahmane
2020 IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing  
Index Terms-Anomalous patterns, change detection (CD), deep learning, feature space reconstruction, heterogeneous remote sensing, multimodal anomaly detector, reconstruction error, sparse autoencoder.  ...  The model starts by learning from image pairs the normal existing patterns in the before and after images to come up with a suitable representation of the normal (nonchange) class.  ...  ACKNOWLEDGMENT The authors would like to acknowledge all other researchers who made at our disposal the CD dataset in order to validate the proposed anomaly CD model.  ... 
doi:10.1109/jstars.2020.2964409 fatcat:j23jop7hmra4zio42bhhculyk4

A survey of non-exchangeable priors for Bayesian nonparametric models [article]

Nicholas J. Foti, Sinead Williamson
2012 arXiv   pre-print
Since the concept of dependent nonparametric processes was formalized by MacEachern [1], there have been a number of models proposed and used in the statistics and machine learning literatures.  ...  Such models are appropriate priors when exchangeability assumptions do not hold, and instead we want our model to vary fluidly with some set of covariates.  ...  The most common Bayesian nonparametric model for these problems is a sparse factor model where y i = Dw i + i , where the columns of D ∈ R m 2 ×∞ are the dictionary elements (or factors), w ∈ R ∞ are the  ... 
arXiv:1211.4798v1 fatcat:asigyslavjaafiqpqov7rx2pyi

Learning to Learn with Compound HD Models

Ruslan Salakhutdinov, Joshua B. Tenenbaum, Antonio Torralba
2011 Neural Information Processing Systems  
We present efficient learning and inference algorithms for the HDP-DBM model and show that it is able to learn new concepts from very few examples on CIFAR-100 object recognition, handwritten character  ...  We introduce HD (or "Hierarchical-Deep") models, a new compositional learning architecture that integrates deep learning models with structured hierarchical Bayesian models.  ...  The proposed model allows for both: a nonparametric prior over potentially unbounded number of global topics, or higher-level features, as well as a nonparametric prior that allow learning an arbitrary  ... 
dblp:conf/nips/SalakhutdinovTT11 fatcat:zgmboicurrcntkx2hih33ylhmm

Special Section Guest Editorial: Feature and Deep Learning in Remote Sensing Applications

John E. Ball, Derek T. Anderson, Chee Seng Chan
2018 Journal of Applied Remote Sensing  
A common theme encountered was the use of nonremote sensing pretrained networks and transfer learning.  ...  Most articles used or extended convolutional neural networks (CNNs) and were application oriented, with a few providing new deep learning models and modules.  ...  Abdi et al. in "Spectral-spatial feature learning for hyperspectral imagery classification using deep stacked sparse autoencoder" uses an unsupervised stacked sparse autoencoder to extract high-level feature  ... 
doi:10.1117/1.jrs.11.042601 fatcat:pq3xg2sggfdtljjs3hrmp7tzdm

Bayesian Multi-label Learning with Sparse Features and Labels, and Label Co-occurrences

He Zhao, Piyush Rai, Lan Du, Wray L. Buntine
2018 International Conference on Artificial Intelligence and Statistics  
We present a probabilistic, fully Bayesian framework for multi-label learning.  ...  The proposed framework has the following appealing aspects: (1) It leverages the sparsity in the label matrix and the feature matrix, which results in very efficient inference, especially for sparse datasets  ...  Our latent factor model for the label co-occurrence is learned jointly with the latent factor model for the label matrix, and sharing the latent factors of the label helps in effectively transferring information  ... 
dblp:conf/aistats/ZhaoRDB18 fatcat:lyzw2ndysvdcvjhlqhxwxb3zyy

Statistical Methods with Applications in Data Mining: A Review of the Most Recent Works

Joaquim Fernando Pinto da Pinto da Costa, Manuel Cabral
2022 Mathematics  
For this reason, we have seen considerable advances over the past few years in statistical methods in data mining.  ...  In such cases, knowledge transfer or transfer learning between task domains would be desirable.  ...  Decision trees (DT) are a nonparametric supervised learning method used for classification and regression.  ... 
doi:10.3390/math10060993 fatcat:j5rz75qv6nburpq5rvsddw3cmu

Learning with Hierarchical-Deep Models

R. Salakhutdinov, J. B. Tenenbaum, A. Torralba
2013 IEEE Transactions on Pattern Analysis and Machine Intelligence  
We present efficient learning and inference algorithms for the HDP-DBM model and show that it is able to learn new concepts from very few examples on CIFAR-100 object recognition, handwritten character  ...  We introduce HD (or "Hierarchical-Deep") models, a new compositional learning architecture that integrates deep learning models with structured hierarchical Bayesian (HB) models.  ...  The proposed model allows for both: a nonparametric prior over potentially unbounded number of global topics, or higher level features, as well as a nonparametric prior that allows learning an arbitrary  ... 
doi:10.1109/tpami.2012.269 pmid:23787346 fatcat:ecdmqr225nfati75px3fbcocaa

Nonparametric Scene Parsing via Label Transfer

Ce Liu, J. Yuen, A. Torralba
2011 IEEE Transactions on Pattern Analysis and Machine Intelligence  
In this paper, we propose a novel, nonparametric approach for object recognition and scene parsing using a new technology we name label transfer.  ...  While there has been a lot of recent work on object recognition and image understanding, the focus has been on carefully establishing mathematical models for images, scenes, and objects.  ...  At a higher level, we can also distinguish two types of object recognition approaches: parametric approaches that consist of learning generative/discriminative models, and nonparametric approaches that  ... 
doi:10.1109/tpami.2011.131 pmid:21709305 fatcat:wcarzzyegvagva6p3uefwjccki
« Previous Showing results 1 — 15 out of 2,636 results