• Corpus ID: 1252271

Learning Modular Structures from Network Data and Node Variables

@inproceedings{Azizi2014LearningMS,
  title={Learning Modular Structures from Network Data and Node Variables},
  author={Elham Azizi and Edoardo M. Airoldi and James E. Galagan},
  booktitle={International Conference on Machine Learning},
  year={2014}
}
A standard technique for understanding underlying dependency structures among a set of variables posits a shared conditional probability distribution for the variables measured on individuals within a group. This approach is often referred to as module networks, where individuals are represented by nodes in a network, groups are termed modules, and the focus is on estimating the network structure among modules. However, estimation solely from node-specific variables can lead to spurious… 

Figures and Tables from this paper

Parallel Construction of Module Networks

This paper presents the first scalable distributed-memory parallel solution for constructing MoNets by parallelizing Lemon-Tree, a widely used sequential software and demonstrates the scalability of this parallel method on a key application of MoNet - the construction of genome-scale gene regulatory networks.

A Nonparametric Multi-view Model for Estimating Cell Type-Specific Gene Regulatory Networks

A Bayesian hierarchical multi-view mixture model termed Symphony that simultaneously learns clusters of cells representing cell types and their underlying gene regulatory networks by integrating data from two views, by explaining gene-gene covariances with the biological machinery regulating gene expression.

Scalable Model Selection for Large-Scale Factorial Relational Models

Stochastic factorized asymptotic Bayesian (sFAB) inference is proposed, a highly-efficient algorithm having both scalability and an inherent model selection capability in a single inference framework that combines concepts in two recently-developed techniques: stochastic variational inference (SVI) and FAB inference.

Mining combined causes in large data sets

Mining Combined Causes

This paper proposes a novel approach to address this practical causal discovery problem, i.e. mining combined causes in large data sets with a high computational efficiency.

Learning mixed membership models with a separable latent structure: Theory, provably efficient algorithms, and applications

In a wide spectrum of problems in science and engineering that includes hyperspectral imaging, gene expression analysis, and machine learning tasks such as topic modeling, the observed data is

Visual Causality Analysis of Event Sequence Data

A visual analytics method for recovering causalities in event sequence data is introduced and the Granger causality analysis algorithm is extended to incorporate user feedback into causal model refinement to demonstrate the usefulness of the system.

A I ] 1 5 O ct 2 01 5 Mining Combined Causes in Large Data Sets

This paper proposes a novel approach to address this practical causal discovery problem, i.e. mining combined causes in large data sets with a high computational efficiency.

An Online Anomaly Learning and Forecasting Model for Large-Scale Service of Internet of Things

  • Junping WangShihui Duan
  • Computer Science
    2014 International Conference on Identification, Information and Knowledge in the Internet of Things
  • 2014
An online anomaly learning and forecasting mechanism for large-scale service of Internet of Thing using the reversible-jump MCMC learning to online learn anomaly-free of dynamics network and service data and performs a structural analysis of IoT-based service topology by Network Utility Maximization theory.

A new online anomaly learning and detection for large-scale service of Internet of Thing

This paper presents a new online anomaly learning and detection mechanism for large-scale service of Internet of Thing that uses the reversible-jump MCMC learning to online learn anomaly-free of dynamics network and service data.

References

SHOWING 1-10 OF 47 REFERENCES

Joint Learning of Modular Structures from Multiple Data Types

An extended model inspired by module networks and stochastic blockmodels for joint learning of structures from observed object variables and relational data among objects and a reversible-jump MCMC learning procedure for learning modules and model parameters is proposed.

Learning Module Networks

Evaluation on real data in the domains of gene expression and the stock market shows that module networks generalize better than Bayesian networks, and that the learned module network structure reveals regularities that are obscured in learnedBayesian networks.

Validating module network learning algorithms using simulated data

The use of the synthetic data generator SynTReN to develop and test an alternative module network learning strategy, which is incorporated in the software package LeMoNe, and evidence is provided that this alternative strategy has several advantages with respect to existing methods is provided.

Module networks revisited: computational assessment and prioritization of model predictions

This work revisits the approach of Segal et al. to infer regulatory modules and their condition-specific regulators from gene expression data and uses a more representative centroid-like solution extracted from an ensemble of possible statistical models to explain the data.

Sparse matrix-variate Gaussian process blockmodels for network modeling

This work proposes a new relational model for network data, Sparse Matrix-variate Gaussian process Blockmodel (SMGB), which generalizes popular bilinear generative models and captures nonlinear network interactions using a matrix-variations process with latent membership variables.

Overlapping stochastic block models with application to the French political blogosphere

This approach allows the vertices to belong to multiple clusters, and, to some extent, generalizes the well-known Stochastic Block Model, and proposes an approximate inference procedure, based on global and local variational techniques.

Mixed Membership Stochastic Blockmodels

This paper describes a latent variable model of such data called the mixed membership stochastic blockmodel, which extends blockmodels for relational data to ones which capture mixed membership latent relational structure, thus providing an object-specific low-dimensional representation.

Stochastic Blockmodels for Directed Graphs

An iterative scaling algorithm is presented for fitting the model parameters by maximum likelihood and blockmodels that are simple extensions of the p 1 model are proposed specifically for such data.

Physical Network Models

A new framework for inferring models of transcriptional regulation, which is based on annotated molecular interaction graphs, is developed and successfully predicts gene knock-out effects with a high degree of accuracy in a cross-validation setting.

Inferring Regulatory Networks from Expression Data Using Tree-Based Methods

This article presents GENIE3, a new algorithm for the inference of GRNs that was best performer in the DREAM4 In Silico Multifactorial challenge and compares favorably with existing algorithms to decipher the genetic regulatory network of Escherichia coli.