5,437 Hits in 11.4 sec

Statistical and Computational Trade-offs in Variational Inference: A Case Study in Inferential Model Selection [article]

Kush Bhatia, Nikki Lijing Kuang, Yi-An Ma, Yixin Wang
2022 arXiv   pre-print
In this work, we study this statistical and computational trade-off in variational inference via a case study in inferential model selection.  ...  We finally demonstrate these statistical and computational trade-offs inference across empirical studies, corroborating the theoretical findings.  ...  This work is supported in part by the National Science Foundation Grants NSF-SCALE MoDL(2134209) and NSF-CCF-2112665 (TILOS), the U.S.  ... 
arXiv:2207.11208v1 fatcat:o42bmco6gnc63mxgjbsbrqwrem

Applied inference

Benjamin C. Lee, David Brooks
2010 ACM Transactions on Architecture and Code Optimization (TACO)  
Collectively these studies demonstrate regression models' ability to expose trends and identify optima in diverse design regions, motivating the application of such models in statistical inference for  ...  This paradigm enables more comprehensive design studies by combining spatial sampling and statistical inference.  ...  CONCLUSIONS AND FUTURE DIRECTIONS This article presents the case for applied statistical inference in microarchitectural design, proposing a simulation paradigm that (1) defines a comprehensive design  ... 
doi:10.1145/1839667.1839670 fatcat:4rwuf5ehqjeitheo6n5e6ecbbm

Multimodel Inference

Kenneth P. Burnham, David R. Anderson
2004 Sociological Methods & Research  
Model selection should be based on a well-justified criterion of what is the "best" model, and that criterion should be based on a philosophy about models and model-based statistical inference, including  ...  There is a clear philosophy, a sound criterion based in information theory, and a rigorous statistical foundation for AIC.  ...  Model selection (variable selection in regression is a special case) is a bias versus variance trade-off, and this is the statistical principle of parsimony.  ... 
doi:10.1177/0049124104268644 fatcat:h44lxx2xdrhrjg2sc67k5psmsa

Exploration, inference and prediction in neuroscience and biomedicine [article]

Danilo Bzdok, John Ioannidis
2019 arXiv   pre-print
In this article, we detail the antagonistic philosophies behind two quantitative approaches: certifying robust effects in understandable variables, and evaluating how accurately a built model can forecast  ...  New ways for generating massive data fueled tension between the traditional methodology, used to infer statistically relevant effects in carefully-chosen variables, and pattern-learning algorithms, used  ...  Consequently, inferential data analysis becomes hard if the statistical model is a black box.  ... 
arXiv:1903.10310v1 fatcat:qqvprhlkvngtnpgvurzufreafq

Inference and Prediction Diverge in Biomedicine

Danilo Bzdok, Denis Engemann, Bertrand Thirion
2020 Patterns  
The shift causes tension between traditional regression methods used to infer statistically significant group differences and burgeoning predictive analysis tools suited to forecast an individual's future  ...  In the 20th century, many advances in biological knowledge and evidence-based medicine were supported by p values and accompanying methods.  ...  ACKNOWLEDGMENTS We are indebted to Olivier Grisel and Gael Varoquaux for fruitful discussion on the topic (both INRIA Saclay/France).  ... 
doi:10.1016/j.patter.2020.100119 pmid:33294865 pmcid:PMC7691397 fatcat:o42promlb5czdhntw3tck5k34e

Inference for changes in biodiversity [article]

Amy Willis, John Bunge, Thea Whitman
2015 arXiv   pre-print
This permits inference for changes in richness with covariates and also a test for homogeneity.  ...  We demonstrate the methodology under simulation, in a gut microbiome study (testing for a decrease in richness with antibiotics), and in a soil microbiome study (testing for homogeneity of replicates).  ...  traded off at the expense of highly parametrized models [14, 20] .  ... 
arXiv:1506.05710v1 fatcat:rygiajg7pfdodp3x6iz3eseh7u

Inference for SDE Models via Approximate Bayesian Computation

Umberto Picchini
2014 Journal of Computational And Graphical Statistics  
Simulation studies for a pharmacokinetics/pharmacodynamics model and for stochastic chemical reactions are considered and a MATLAB package implementing our ABC-MCMC algorithm is provided.  ...  The relevance of this class of models is growing in many applied research areas and is already a standard tool to model e.g. financial, neuronal and population growth dynamics.  ...  Therefore a trade-off is necessary, namely select a small δ * while retaining a long enough sequence of draws to allow accurate posterior inference.  ... 
doi:10.1080/10618600.2013.866048 fatcat:kfes6lfycbctnk7vvzmk4vnncm

Bayesian and Frequentist Inference for Ecological Inference: The RxC Case

Ori Rosen, Wenxin Jiang, Gary King, Martin A. Tanner
2001 Statistica neerlandica (Print)  
In the ®nal section of the paper we provide an overview of a range of alternative inferential approaches which trade-off computational intensity for statistical ef®ciency.  ...  In this paper we propose Bayesian and frequentist approaches to ecological inference, based on R 3 C contingency tables, including a covariate.  ...  The statistical ®eld dates to OGBURN and GOLTRA (1919) and GEHLKE (1917) , who, in the context of studies of political behavior, ®rst recognized the ecological inference problem, and ROBINSON (1950  ... 
doi:10.1111/1467-9574.00162 fatcat:636pcelsrfclpp6u5hshv7siae

A theory of learning to infer [article]

Ishita Dasgupta, Eric Schulz, Joshua B. Tenenbaum, Samuel J. Gershman
2019 biorxiv/medrxiv   pre-print
The theory also explains a range of related phenomena: memory effects, belief bias, and the structure of response variability in probabilistic reasoning.  ...  By adapting to the query distribution, the recognition model "learns to infer".  ...  understand how learned and memoryless inference strategies interact and trade-off.  ... 
doi:10.1101/644534 fatcat:skrghmeiwrdorkrfotbhx4hvs4

Object Perception as Bayesian Inference

Daniel Kersten, Pascal Mamassian, Alan Yuille
2004 Annual Review of Psychology  
Recent work in Bayesian theories of visual perception has shown how complexity may be managed and ambiguity resolved through the taskdependent, probabilistic integration of prior object knowledge with  ...  Typical images are highly complex because they consist of many objects embedded in background clutter.  ...  In Basic Bayes: The Trade-Off Between Feature Reliability and Priors (below), we review the perceptual consequences of knowledge specified by the prior p(S), the likelihood p(I |S), and the trade-off between  ... 
doi:10.1146/annurev.psych.55.090902.142005 pmid:14744217 fatcat:vqggertz4rdq7g5bjnfkvr6awu

Exploration, Inference, and Prediction in Neuroscience and Biomedicine

Danilo Bzdok, John P.A. Ioannidis
2019 Trends in Neurosciences  
to obtain p-values, the latter with a stronger heritage in computer science [10] [11] [12] .  ...  New ways for generating massive data fueled tension between the traditional methodology, used to infer statistically relevant effects in carefully-chosen variables, and pattern-learning algorithms, used  ...  Consequently, inferential data analysis becomes hard if the statistical model is a black box.  ... 
doi:10.1016/j.tins.2019.02.001 fatcat:4zi2rj2mcfbbxmqijrnxkfwpaa

Towards more accessible conceptions of statistical inference

C. J. Wild, M. Pfannkuch, M. Regan, N. J. Horton
2011 Journal of the Royal Statistical Society: Series A (Statistics in Society)  
There is a compelling case, based on research in statistics education, for first courses in statistical inference to be underpinned by a staged development path.  ...  These build on novel ways of experiencing sampling variation and have intuitive connections to the standard formal methods of making inferences in first university courses in statistics.  ...  This work was partially supported by a grant from New Zealand's 'Teaching and learning research initiative' (  ... 
doi:10.1111/j.1467-985x.2010.00678.x fatcat:ksjmrzl24zac3jvbv6mwfe5boe


Timothy H. Keitt, Dean L. Urban
2005 Ecology  
Understanding of spatial pattern and scale has been identified as a key issue in ecology, yet ecology has traditionally lacked necessary tools for making inference about relationships between scale-specific  ...  We introduce wavelet-coefficient regression, in which the dependent and independent variables are wavelet transformed prior to analysis, as a means to formalize scale-specific relationships in ecological  ...  Dixon and three anonymous reviewers for constructive comments that greatly improved this manuscript.  ... 
doi:10.1890/04-1016 fatcat:cknl5naqf5cglagkmlibwtqctu

Statistical Inference [article]

Konstantin Zuev
2016 arXiv   pre-print
In these notes statistics is viewed as a branch of mathematical engineering, that studies ways of extracting reliable information from limited data for learning, prediction, and decision making in the  ...  In fact, there is a continuous spectrum of attitudes toward statistics ranging from pure theoreticians, proving asymptotic efficiency and searching for most powerful tests, to wild practitioners, blindly  ...  This [note] provides a good trade-off between rigor and readability. 2.  ... 
arXiv:1603.04929v1 fatcat:7eyivrt2k5ak5ctyf4wc5bht2a

Remembrance of Inferences Past [article]

Ishita Dasgupta, Eric Schulz, Noah D. Goodman, Samuel J. Gershman
2017 bioRxiv   pre-print
These findings support the view that the brain trades off accuracy and computational cost, to make efficient use of its limited cognitive resources to approximate probabilistic inference.  ...  Since people often encounter many closely related distributions, selective reuse of computations (amortized inference) is a computationally efficient use of the brain's limited resources.  ...  E.S. was supported by a postdoctoral fellowship from the Harvard Data Science Initiative.  ... 
doi:10.1101/231837 fatcat:kqgca26hivhx7cattbzwtvv6oq
« Previous Showing results 1 — 15 out of 5,437 results