Beta-Negative Binomial Process and Poisson Factor Analysis

Abstract

A beta-negative binomial (BNB) process is proposed, leading to a beta-gamma-Poisson process, which may be viewed as a "multi-scoop" generalization of the beta-Bernoulli process. The BNB process is augmented into a beta-gamma-gamma-Poisson hierarchical structure, and applied as a nonparametric Bayesian prior for an infinite Poisson factor analysis model. A finite approximation for the beta process Levy random measure is constructed for convenient implementation. Efficient MCMC computations are performed with data augmentation and marginalization techniques. Encouraging results are shown on document count matrix factorization.

Appearing in AISTATS 2012 (submitted on Oct. 2011)

References in corpus (3)

• On Smoothing and Inference for Topic Models
• Discrete Component Analysis
• Stochastic expansions using continuous dictionaries: Lévy adaptive regression kernels

Cited by in corpus (59)

• The Emerging Trends of Multi-Label Learning
• Lidar waveform based analysis of depth images constructed using sparse single-photon data
• Scalable Recommendation with Poisson Factorization
• WHAI: Weibull Hybrid Autoencoding Inference for Deep Topic Modeling
• Cluster and Feature Modeling from Combinatorial Stochastic Processes
• Probabilistic Count Matrix Factorization for Single Cell Expression Data Analysis
• Natural-Parameter Networks: A Class of Probabilistic Neural Networks
• Nonparametric Bayesian Factor Analysis for Dynamic Count Matrices
• Zero-Truncated Poisson Tensor Factorization for Massive Binary Tensors
• Multimodal Data Fusion in High-Dimensional Heterogeneous Datasets via Generative Models
• Truncated Random Measures
• The combinatorial structure of beta negative binomial processes
• Dirichlet belief networks for topic structure learning
• Towards Bayesian Deep Learning: A Framework and Some Existing Methods
• A Survey on Bayesian Deep Learning
• Augment-and-Conquer Negative Binomial Processes
• Scalable Bayesian Non-Negative Tensor Factorization for Massive Count Data
• The Poisson Gamma Belief Network
• Gamma Processes, Stick-Breaking, and Variational Inference
• A unifying representation for a class of dependent random measures
• Scaled subordinators and generalizations of the Indian buffet process
• Matching Visual Features to Hierarchical Semantic Topics for Image Paragraph Captioning
• Softplus Regressions and Convex Polytopes
• Fast Simulation of Hyperplane-Truncated Multivariate Normal Distributions
• Beta Process Non-negative Matrix Factorization with Stochastic Structured Mean-Field Variational Inference
• Combinatorial clustering and the beta negative binomial process
• Data Augmentation for Bayesian Deep Learning
• Gibbs-type Indian buffet processes
• A unified construction for series representations and finite approximations of completely random measures
• Feature allocations, probability functions, and paintboxes
• Nonparametric Bayesian Negative Binomial Factor Analysis
• Topic-Guided Abstractive Text Summarization: a Joint Learning Approach
• Independent finite approximations for Bayesian nonparametric inference
• Correlated Random Measures
• Variational Auto-encoder Based Bayesian Poisson Tensor Factorization for Sparse and Imbalanced Count Data
• Priors for Random Count Matrices Derived from a Family of Negative Binomial Processes
• Distributed, partially collapsed MCMC for Bayesian Nonparametrics
• Nonnegative Bayesian nonparametric factor models with completely random measures for community detection
• Sample Size Dependent Species Models
• Deep Autoencoding Topic Model with Scalable Hybrid Bayesian Inference
• Closed-form Marginal Likelihood in Gamma-Poisson Matrix Factorization
• Non-Gaussian Discriminative Factor Models via the Max-Margin Rank-Likelihood
• Slice Sampling for General Completely Random Measures
• A Comparative Study of Gamma Markov Chains for Temporal Non-Negative Matrix Factorization
• Deep Poisson gamma dynamical systems
• BNP-Seq: Bayesian Nonparametric Differential Expression Analysis of Sequencing Count Data
• Posterior distributions for Hierarchical Spike and Slab Indian Buffet processes
• Gamma Belief Networks
• Statistical inference on errorfully observed graphs
• BayCount: A Bayesian Decomposition Method for Inferring Tumor Heterogeneity using RNA-Seq Counts
• Variational Hetero-Encoder Randomized GANs for Joint Image-Text Modeling
• Black-box constructions for exchangeable sequences of random multisets
• On collapsed representation of hierarchical Completely Random Measures
• A survey of non-exchangeable priors for Bayesian nonparametric models
• Generalized Negative Binomial Processes and the Representation of Cluster Structures
• Generalized Bregman Divergence and Gradient of Mutual Information for Vector Poisson Channels
• Hierarchical Topic Presence Models
• GO Hessian for Expectation-Based Objectives
• Discrete factor analysis