Filters








12,892 Hits in 2.9 sec

Probabilistic latent maximal marginal relevance

Shengbo Guo, Scott Sanner
2010 Proceeding of the 33rd international ACM SIGIR conference on Research and development in information retrieval - SIGIR '10  
Perhaps one of the most well-known and most used algorithms for result set diversication is that of Maximal Marginal Relevance (MMR).  ...  This novel derivation presents a formal probabilistic latent view of MMR (PLMMR) that (a) removes the need to manually balance relevance and diversity parameters, (b) shows that specific definitions of  ...  PROBABILISTIC LATENT MMR w 1 r 1 w 1 r 2 t' w 1 s 1 w 1 s 2 w 1 t 1 w 1 t 2 ... q We begin our discussion of PLMMR by introducing a graphical model of (marginal) relevance in Figure 1 .  ... 
doi:10.1145/1835449.1835639 dblp:conf/sigir/GuoS10 fatcat:cp6a2oal2fgcrmqfktxto3ljui

Manifold Relevance Determination: Learning the Latent Space of Robotics [article]

Pete Trautman
2017 arXiv   pre-print
In this article we present the basics of manifold relevance determination (MRD) as introduced in mrd, and some applications where the technology might be of particular use.  ...  [Tipping and Bishop, 1999] ), reduces to PCA if one chooses the parameters W that maximize the marginal likelihood W * = arg max W p(Y | W, β).  ...  Finally, MRD introduces an approximate latent space marginalization which in turn allows for automatic relevance determination priors to be used, thus enabling soft boundaries between shared and private  ... 
arXiv:1705.03158v2 fatcat:jukr7cx6vvhtrikemqgq6zg7di

Probabilistic Factorization of Non-negative Data with Entropic Co-occurrence Constraints [chapter]

Paris Smaragdis, Madhusudana Shashanka, Bhiksha Raj, Gautham J. Mysore
2009 Lecture Notes in Computer Science  
In this paper we present a probabilistic algorithm which factorizes non-negative data.  ...  We employ entropic priors to additionally satisfy that user specified pairs of factors in this model will have their cross entropy maximized or minimized.  ...  We recast the task of non-negative factorization as a probabilistic latent variable decomposition on count/histogram data.  ... 
doi:10.1007/978-3-642-00599-2_42 fatcat:mybpyx7e2vhpvepcy3nbo2jtrm

A Latent Variable Recurrent Neural Network for Discourse Relation Language Models [article]

Yangfeng Ji and Gholamreza Haffari and Jacob Eisenstein
2016 arXiv   pre-print
The discourse relations are represented with a latent variable, which can be predicted or marginalized, depending on the task.  ...  Furthermore, by marginalizing over latent discourse relations at test time, we obtain a discourse informed language model, which improves over a strong LSTM baseline.  ...  In contrast, the discrete latent variables in our model are easy to sum and maximize over.  ... 
arXiv:1603.01913v2 fatcat:4bsf55cb6rfxrnivsqwynenu3u

A Latent Variable Recurrent Neural Network for Discourse-Driven Language Models

Yangfeng Ji, Gholamreza Haffari, Jacob Eisenstein
2016 Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies  
The discourse relations are represented with a latent variable, which can be predicted or marginalized, depending on the task.  ...  Furthermore, by marginalizing over latent discourse relations at test time, we obtain a discourse informed language model, which improves over a strong LSTM baseline.  ...  In contrast, the discrete latent variables in our model are easy to sum and maximize over.  ... 
doi:10.18653/v1/n16-1037 dblp:conf/naacl/JiHE16 fatcat:e3l2es32fjhxpgh2g4qdfse5za

Learning Similarity with Probabilistic Latent Semantic Analysis for Image Retrieval

2015 KSII Transactions on Internet and Information Systems  
In this paper, we propose a similarity learning method on the basis of probabilistic generative model, i.e., probabilistic latent semantic analysis (PLSA).  ...  Then, the parameters are determined through simultaneously maximizing the log likelihood function of PLSA and the retrieval performance over the training dataset.  ...  Latent Semantic Analysis In this work, we utilize Probabilistic Latent Semantic Analysis (PLSA) [29] to model the distribution of images represented in bag of visual words quantized from image features  ... 
doi:10.3837/tiis.2015.04.009 fatcat:vc3po2xtqrdbbdom4mjb5xkf64

Generative Models that Discover Dependencies Between Data Sets

Arto Klami, Samuel Kaski
2006 Machine Learning for Signal Processing  
We develop models for a kind of data fusion task: Combine multiple data sources under the assumption that data set specific variation is irrelevant and only between-data variation is relevant.  ...  From this perspective the proposed clustering method is a probabilistic alternative to AC that directly maximizes a measure of dependency.  ...  CCA and marginal model complexity In Section 3 we claimed that the probabilistic formulation of CCA only holds when sufficiently complex marginal models are used.  ... 
doi:10.1109/mlsp.2006.275534 fatcat:3rbcvvyzozgedf25rz7hduuoxa

Probabilistic approach to detecting dependencies between data sets

Arto Klami, Samuel Kaski
2008 Neurocomputing  
We study data fusion under the assumption that data source-specific variation is irrelevant and only shared variation is relevant.  ...  Traditionally the shared variation has been sought by maximizing a dependency measure, such as correlation of linear projections in Canonical Correlation Analysis.  ...  The number of clusters was chosen so that the lower bound for the marginalized likelihood (4) was maximized.  ... 
doi:10.1016/j.neucom.2007.12.044 fatcat:eho75neeazgcxn4leqrpfim45q

Variational Mutual Information Maximization Framework for VAE Latent Codes with Continuous and Discrete Priors [article]

Andriy Serdega, Dae-Shik Kim
2020 arXiv   pre-print
In comparison to other methods, it provides an explicit objective that maximizes lower bound on mutual information between latent codes and observations.  ...  We propose Variational Mutual Information Maximization Framework for VAE to address this issue.  ...  The resulting directed probabilistic model is depicted in In this setting, we again wish to maximize marginal data log-likelihood log p θ (x) by maximizing its variational lower bound.  ... 
arXiv:2006.02227v1 fatcat:trduf6cokfh5xeqx3t3xnqlnfi

Latent trees for estimating intensity of Facial Action Units

Sebastian Kaltwang, Sinisa Todorovic, Maja Pantic
2015 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
A graph-edit operation that increases maximally the likelihood and minimally the model complexity is selected as optimal in each iteration.  ...  Effectiveness of our structure learning is demonstrated by probabilistically sampling meaningful facial expressions from the LT.  ...  A graph-edit operation that increases maximally the likelihood and minimally the model complexity is selected as optimal in each iteration.  ... 
doi:10.1109/cvpr.2015.7298626 dblp:conf/cvpr/KaltwangTP15 fatcat:n2s3jyxilzfbvlhgqnyrz22nye

Getting Started in Probabilistic Graphical Models

Edoardo M. Airoldi
2007 PLoS Computational Biology  
How can we use PGMs to discover patterns that are biologically relevant? And to what extent can PGMs help us formulate new hypotheses that are testable at the bench?  ...  Probabilistic graphical models (PGMs) have become a popular tool for computational analysis of biological data in a variety of domains. But, what exactly are they and how do they work?  ...  Assessment of biological relevance can be qualitative or quantitative.  ... 
doi:10.1371/journal.pcbi.0030252 pmid:18069887 pmcid:PMC2134967 fatcat:ja62f3h3tnbpfmmelkntaqr624

Elementary Sources: Latent Component Analysis For Music Composition

Spencer S. Topel, Michael A. Casey
2011 Zenodo  
Figure 3 illustrates the decomposition of magnitude STFT time-frequency distributions into probabilistic latent components using this 2-dimensional marginal decomposition algorithm.  ...  these steps converges to a solution for the marginals and the latent variable priors.  ... 
doi:10.5281/zenodo.1417350 fatcat:epa3fnkhsvbnjjei5jq3im4hva

Probabilistic Classification Vector Machines

Huanhuan Chen, P. Tino, Xin Yao
2009 IEEE Transactions on Neural Networks  
We compare PCVMs with soft-margin support vector machines (SVM Soft ), hard-margin support vector machines (SVM Hard ), SVM with the kernel parameters optimized by PCVMs (SVM PCVM ), relevance vector machines  ...  In this paper, a sparse learning algorithm, probabilistic classification vector machines (PCVMs), is proposed.  ...  The success of SVMs is attributed to the margin maximization theory [27] .  ... 
doi:10.1109/tnn.2009.2014161 pmid:19398403 fatcat:ky2h4o7emzaxjog3x4cyl4c2gi

Dependency detection with similarity constraints [article]

Leo Lahti, Samuel Myllykangas, Sakari Knuutila, Samuel Kaski
2011 arXiv   pre-print
CCA searches for a linear projection for each view, such that the correlations between the projections are maximized.  ...  The simplified model does not distinguish between the shared and marginal effects as effectively as the full probabilistic CCA but it has fewer model parameters.  ...  We suggest also a probabilistic version for constrained dependency search that provides a robust alternative for direct maximization of correlations.  ... 
arXiv:1101.5919v1 fatcat:4mnjxltdofffzome7fyg6iwk7u

Probabilistic latent query analysis for combining multiple retrieval sources

Rong Yan, Alexander G. Hauptmann
2006 Proceedings of the 29th annual international ACM SIGIR conference on Research and development in information retrieval - SIGIR '06  
To merge retrieval sources adaptively according to query topics, we propose a series of new approaches called probabilistic latent query analysis (pLQA), which can associate non-identical combination weights  ...  with latent classes underlying the query space.  ...  margin relevance.  ... 
doi:10.1145/1148170.1148228 dblp:conf/sigir/YanH06 fatcat:r2cri2qq4bg37jl4e3ecq2lsn4
« Previous Showing results 1 — 15 out of 12,892 results