Accelerated Gibbs Sampling for Infinite Sparse Factor Analysis

Andrzejewski, David

doi:10.2172/1026471

Cited by 4 publications

(5 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…One limitation of this approach is its computational cost. The complexity periteration of the IBP sampler is O(N 3 (K 2 + KP )) due to the matrix in the exponent of the collapsed likelihood (23). This motivates next section which derives an accelerated Gibbs sampling following ideas from [15].…”

Section: Collapsed Gibbs Sampling For Inpaintingmentioning

confidence: 96%

See 1 more Smart Citation

Indian Buffet Process dictionary learning: Algorithms and applications to image processing

Dang

Chainais

2017

International Journal of Approximate Reasoning

View full text Add to dashboard Cite

Ill-posed inverse problems call for some prior model to define a suitable set of solutions. A wide family of approaches relies on the use of sparse representations. Dictionary learning precisely permits to learn a redundant set of atoms to represent the data in a sparse manner. Various approaches have been proposed, mostly based on optimization methods. We propose a Bayesian non parametric approach called IBP-DL that uses an Indian Buffet Process prior. This method yields an efficient dictionary with an adaptive number of atoms. Moreover the noise and sparsity levels are also inferred so that no parameter tuning is needed. We elaborate on the IBP-DL model to propose a model for linear inverse problems such as inpainting and compressive sensing beyond basic denoising. We derive a collapsed and an accelerated Gibbs samplers and propose a marginal maximum a posteriori estimator of the dictionary. Several image processing experiments are presented and compared to other approaches for illustration.

show abstract

Section: Collapsed Gibbs Sampling For Inpaintingmentioning

confidence: 96%

“…This is the reason why we have followed the recommendations in [15] to perform an accelerated Gibbs sampling. The two methods and their complexity were compared in [23]. A study comparing the speed of different Gibbs samplers for Z when {H i } are identity matrices can be found in [15].…”

Section: Accelerated Gibbs Sampling For Inpaintingmentioning

confidence: 99%

Indian Buffet Process dictionary learning: Algorithms and applications to image processing

Dang

Chainais

2017

International Journal of Approximate Reasoning

View full text Add to dashboard Cite

show abstract

“…Rather, results are accompanied by estimates of uncertainty, which are conditioned on the available data. However, this sample is larger than other studies using NKI-RS data to investigate risky and impulsive behavior, including substance use 38,39 .…”

Section: Methods Participantsmentioning

confidence: 97%

Distinct Neurocognitive Fingerprints Reflect Differential Associations with Risky and Impulsive Behavior in a Neurotypical Sample

Ruiz¹,

Brazil²,

Baskin–Sommers³

2022

Preprint

View full text Add to dashboard Cite

Research on neurocognitive deficits related to risky and impulsive behaviors has been limited by analytic approaches that treat neurocognitive components as functionally dissociable or that fail to adequately capture individual variability in function. In a neurotypical sample (N=673), we characterized neurocognition via a novel application of a Bayesian latent feature learning model – the Indian Buffet Process. All features were within a relatively normative range of neurocognition; however, subtle variability related to risky and impulsive behaviors. The relatively overall poorer neurocognition feature correlated with greater affective impulsivity and substance use patterns/problems. The poorer episodic memory and emotion feature correlated with greater trait externalizing and sensation-seeking. The poorer attention feature correlated with increased trait externalizing and negative urgency, but decreased positive urgency and substance use. Finally, average or mixed features negatively correlated with various outcomes. Estimating nuanced patterns of neurocognition can inform our understanding of a continuum of risky and impulsive behaviors.

show abstract

“…Completeness for each of the 27 neurocognitive variables, impulsive behavior, and substance use outcomes is documented in Table S1. The Gibbs sampling procedure also helps the model converge with fewer iterations 36,41 .…”

Section: Methodsmentioning

confidence: 99%

Distinct neurocognitive fingerprints reflect differential associations with risky and impulsive behavior in a neurotypical sample

Ruiz

Brazil

Baskin‐Sommers

2023

Sci Rep

View full text Add to dashboard Cite

Engagement in risky and impulsive behavior has long been associated with deficits in neurocognition. However, we have a limited understanding of how multiple subfunctions of neurocognition co-occur within individuals and which combinations of neurocognitive subfunctions are most relevant for risky and impulsive behavior. Using the neurotypical Nathan Kline Institute Rockland Sample (N = 673), we applied a Bayesian latent feature learning model—the Indian Buffet Process—to identify nuanced, individual-specific profiles of multiple neurocognitive subfunctions and examine their relationship to risky and impulsive behavior. All features were within a relatively normative range of neurocognition; however, there was subtle variability related to risky and impulsive behaviors. The relatively overall poorer neurocognition feature correlated with greater affective impulsivity and substance use patterns/problems. The poorer episodic memory and emotion feature correlated with greater trait externalizing and sensation-seeking. The poorer attention feature correlated with increased trait externalizing and negative urgency but decreased positive urgency and substance use. Finally, the average or mixed features negatively correlated with various risky and impulsive behaviors. Estimating nuanced patterns of co-occurring neurocognitive functions can inform our understanding of a continuum of risky and impulsive behaviors.

show abstract

Accelerated Gibbs Sampling for Infinite Sparse Factor Analysis

Cited by 4 publications

References 4 publications

Indian Buffet Process dictionary learning: Algorithms and applications to image processing

Indian Buffet Process dictionary learning: Algorithms and applications to image processing

Distinct Neurocognitive Fingerprints Reflect Differential Associations with Risky and Impulsive Behavior in a Neurotypical Sample

Distinct neurocognitive fingerprints reflect differential associations with risky and impulsive behavior in a neurotypical sample

Contact Info

Product

Resources

About