16,336 Hits in 5.1 sec

Implicit Variational Inference: the Parameter and the Predictor Space [article]

Yann Pequignot, Mathieu Alain, Patrick Dallaire, Alireza Yeganehparast, Pascal Germain, Josée Desharnais, François Laviolette
2020 arXiv   pre-print
While most methods perform inference in the space of parameters, we explore the benefits of carrying inference directly in the space of predictors.  ...  Relying on a family of distributions given by a deep generative neural network, we present two ways of carrying variational inference: one in parameter space, one in predictor space.  ...  114090, and the Canada CIFAR AI Chair Program.  ... 
arXiv:2010.12995v1 fatcat:pdgx3mgeozhubgonh3afdicjl4

Implicit House Prices: Variation over Time and Space in Spain

Stanley McGreal, Paloma Taltavull de La Paz
2013 Urban Studies  
The paper employs a large database from the Spanish housing market with models generated to explain how the pricing of attributes varies by region and how variation over time impacts on explanatory power  ...  It is shown that clustering by time tends to give higher parameter values and places more relevance on neighbourhood values.  ...  The analysis isolates the space effect to allow for the different provinces and time effects to capture annual variation.  ... 
doi:10.1177/0042098012471978 fatcat:4mn7jm5k65cdbfpnhuhc6ows4q

Latent space projection predictive inference [article]

Alejandro Catalina, Paul Bürkner, Aki Vehtari
2021 arXiv   pre-print
We extend projection predictive inference to enable computationally efficient variable and structure selection in models outside the exponential family.  ...  By adopting a latent space projection predictive perspective we are able to: 1) propose a unified and general framework to do variable selection in complex models while fully honouring the original model  ...  Acknowledgments The authors thank Aalto Science-IT for computational resources and FCAI for funding and support.  ... 
arXiv:2109.04702v1 fatcat:4ywvp3cdkjf23ktoihlipnyedm

Multimodel Inference

Kenneth P. Burnham, David R. Anderson
2004 Sociological Methods & Research  
Such a framework and methodology allows us to go beyond inference based on only the selected best model. Rather, we do inference based on the full set of models: multimodel inference.  ...  AIC can be justified as Bayesian using a "savvy" prior on models that is a function of sample size and the number of model parameters. Furthermore, BIC can be derived as a non-Bayesian result.  ...  This is a likelihood function over the model set in the sense that L(θ|data, g i ) is the likelihood over the parameter space (for model g i ) of the parameter θ, given the data (x) and the model (g i  ... 
doi:10.1177/0049124104268644 fatcat:h44lxx2xdrhrjg2sc67k5psmsa

Advances in Variational Inference [article]

Cheng Zhang, Judith Butepage, Hedvig Kjellstrom, Stephan Mandt
2018 arXiv   pre-print
with atypical divergences, and (d) amortized VI, which implements the inference over local latent variables with inference networks.  ...  These models are usually intractable and thus require approximate inference.  ...  ACKNOWLEDGMENTS The authors would like to thank Yingzhen Li, Sebastian Nowozin, Francisco Ruiz, Tianfan Fu, Robert Bamler, and especially Andrew Hartnett for comments and discussions that greatly improved  ... 
arXiv:1711.05597v3 fatcat:st53lmyx5ndpvezmw6vhw4fnhy

Advances in Variational Inference

Cheng Zhang, Judith Butepage, Hedvig Kjellstrom, Stephan Mandt
2018 IEEE Transactions on Pattern Analysis and Machine Intelligence  
with atypical divergences, and (d) amortized VI, which implements the inference over local latent variables with inference networks.  ...  These models are usually intractable and thus require approximate inference.  ...  ACKNOWLEDGMENTS The authors would like to thank Sebastian Nowozin, Francisco Ruiz, Tianfan Fu, Robert Bamler, and especially Andrew Hartnett and Yingzhen Li for comments and discussions that greatly improved  ... 
doi:10.1109/tpami.2018.2889774 fatcat:xffyfbw5w5c4dklgs3uvwynp3u

Implicit Bayesian Inference Using Option Prices

Gael M. Martin, Catherine S. Forbes, Vance L. Martin
2005 Journal of Time Series Analysis  
A range of models allowing for conditional leptokurtosis, skewness and time-varying volatility in returns are considered, with posterior parameter distributions and model probabilities backed out from  ...  A Bayesian approach to option pricing is presented, in which posterior inference about the underlying returns process is conducted implicitly via observed option prices.  ...  in (15) to (17) (15) and (17) , again at each point in the parameter space and at each point in (simulated) time.  ... 
doi:10.1111/j.1467-9892.2005.00410.x fatcat:kchenukkpjgenbp4wd6wr6v5hm

Network inference and biological dynamics

Chris. J. Oates, Sach Mukherjee
2012 Annals of Applied Statistics  
Our investigation sheds light on the applicability and limitations of network inference and provides guidance for practitioners and suggestions for experimental design.  ...  This clarifies the link between biochemical networks as they operate at the cellular level and network inference as carried out on data that are averages over populations of cells.  ...  Kafadar and the anonymous referees for constructive suggestions that helped to improve the content and presentation of this article, and G. O. Roberts, S. Spencer and S. M.  ... 
doi:10.1214/11-aoas532 pmid:23284600 pmcid:PMC3533376 fatcat:ogs4kpfyebamrbrotoqw5z5wze

Geostatistical inference under preferential sampling

Peter J. Diggle, Raquel Menezes, Ting-li Su
2010 Journal of the Royal Statistical Society, Series C: Applied Statistics  
Preferential sampling arises when the process that determines the data-locations and the process being modelled are stochastically dependent.  ...  We present a model for preferential sampling, and demonstrate through simulated examples that ignoring preferential sampling can lead to seriously misleading inferences.  ...  the parameter space taken by the various runs of the Nelder-Mead optimisation algorithm.  ... 
doi:10.1111/j.1467-9876.2009.00701.x fatcat:xelkbz7pjzap7ml7lxrixxaz3u

Bayesian and Frequentist Inference for Synthetic Controls [article]

Ignacio Martinez, Jaume Vives-i-Bastida
2022 arXiv   pre-print
treated unit and that our Bayes estimator is asymptotically close to the MLE in the total variation sense.  ...  Then, we propose a Bayesian alternative to the synthetic control method that preserves the main features of the standard method and provides a new way of doing valid inference.  ...  However, as discussed, our identification result requires that the donor pool grows with the sample size, J → ∞. This means that the parameter space is growing with the sample size T 0 .  ... 
arXiv:2206.01779v1 fatcat:545tjebucjeojjcb63e5fjtige

Inferring Network Structure From Data [article]

Ivan Brugere, Tanya Y. Berger-Wolf
2020 arXiv   pre-print
Yet, the impact of the various choices in translating this data to a network have been largely unexamined.  ...  We demonstrate that this network definition matters in several ways for modeling the behavior of the underlying system.  ...  Methods Model selection for task-focused network inference selects a network model and associated parameters of structured predictors to best perform a task or set of tasks on the topology.  ... 
arXiv:2004.02046v1 fatcat:5rexqmojxbbmpauach52lqulsq

Revisiting Causality Inference in Memory-less Transition Networks [article]

Abbas Shojaee, Isuru Ranasinghe, Alireza Ani
2016 arXiv   pre-print
We show that CICT is highly accurate in inferring whether the transition between a pair of events is causal or random and performs well in identifying the direction of causality in a bi-directional association  ...  We applied machine learning models to learn these different behaviors and to infer causality. We name this new method Causality Inference using Composition of Transitions (CICT).  ...  For nonlinear systems, different methods have been applied including nonlinear variations of Granger causality 3, 4 , techniques of state space reconstruction 2,5,6 , conditional mutual information  ... 
arXiv:1608.02658v3 fatcat:hckfxgthj5djdmsoavdrxlic54

An Explanation of In-context Learning as Implicit Bayesian Inference [article]

Sang Michael Xie, Aditi Raghunathan, Percy Liang, Tengyu Ma
2022 arXiv   pre-print
Here, the LM must infer a latent document-level concept to generate coherent next tokens during pretraining.  ...  At test time, in-context learning occurs when the LM also infers a shared latent concept between examples in a prompt.  ...  Acknowledgements We thank Tianyi Zhang, Frieda Rong, Lisa Li, Colin Wei, Shibani Santurkar, Tri Dao, Ananya Kumar, and  ... 
arXiv:2111.02080v6 fatcat:iglwaptkqbhd7nw3jopzq63vxy

Fast and Accurate Estimation of Non-Nested Binomial Hierarchical Models Using Variational Inference [article]

Max Goplerud
2021 arXiv   pre-print
However, inference in the presence of non-nested effects and on large datasets is challenging and computationally burdensome. This paper provides two contributions to scalable and accurate inference.  ...  Second, I propose "marginally augmented variational Bayes" (MAVB) that further improves the initial approximation through a step of Bayesian post-processing.  ...  "Semi-Implicit Variational Inference." In International Conference on Machine Learning. Zhao, Y., Staudenmayer, J., Coull, B. A., and Wand, M. P. (2006).  ... 
arXiv:2007.12300v3 fatcat:rmp4pej2cfayxd4olwmmhgu3pq

Bayesian Neural Network Inference via Implicit Models and the Posterior Predictive Distribution [article]

Joel Janek Dabrowski, Daniel Edward Pagendam
2022 arXiv   pre-print
The approach is more scalable to large data than Markov Chain Monte Carlo, it embraces more expressive models than Variational Inference, and it does not rely on adversarial training (or density ratio  ...  approximate posterior distribution over the parameters of the primary model.  ...  Acknowledgements We would like to thank Edwin Bonilla for the various discussions around this topic of this paper. This work was supported by the CSIRO MLAI Future Science Platform.  ... 
arXiv:2209.02188v1 fatcat:ps57mraoqnfhvht7zoiksq3em4
« Previous Showing results 1 — 15 out of 16,336 results